A Symposium to Celebrate the 25th Anniversary of the PAH ...

EAS Publications Series, Volume 46, 2011

PAHs and the Universe: A Symposium toCelebrate the 25th Anniversary of the PAH

Hypothesis

Toulouse, France

31 May - 04 June, 2010

Edited by: C. Joblin and A.G.G.M. Tielens

17 avenue du Hoggar, PA de Courtabœuf, B.P. 112, 91944 Les Ulis cedex A, France

First pages of all issues in the series and full-text articles

in PDF format are available to registered users at:

http://www.eas-journal.org

Sponsors

Centre National de la Recherche Scientifique (CNRS-INSU, INP, INC)

Universite Toulouse III - Paul Sabatier (UPS)

Centre National d’Etudes Spatiales (CNES)

Region Midi-Pyrenees

Ministere de l’Enseignement Superieur et de la Recherche

European Space Agency (ESA)

IRSAMC-UPSReseau de Chimie Theorique

Departement de la Haute Garonne

Scientific Organizing Committee

V. Bierbaum, University of Boulder, USA

L. d’Hendecourt, IAS, Paris, France

T. Geballe, Gemini Observatory, Hawaii, USA

K. Gordon, STScI, Baltimore, USA

T. Henning, MPIA, Heidelberg, Germany

E. Herbst, Ohio State University, USA

C. Joblin (co-chair), CESR, Toulouse, France

R. Kennicutt, University of Cambridge, USA

D. Lutz, MPE, Garching, Germany

T. Onaka, University of Tokyo, Japan

J.-L. Puget, IAS, Paris, France

F. Salama, NASA ARC, USA

S. Schlemmer, University of Cologne, Germany

F. Spiegelman, LCPQ, Toulouse, France

A. Tielens (co-chair), Leiden Observatory, The Netherlands

Cover Figure

New light on and from PAHs: From Orion to Le Pont de Toulouse. Image credits:

Orion as seen by Spitzer, NASA/JPL-Caltech/S.T. Megeath (University of Toledo, Ohio);

Toulouse, L. Montier.

Indexed in: ADS, Current Contents Proceedings – Engineering & Physical Sciences,

ISTP r©/ISI Proceedings, ISTP/ISI CDROM Proceedings.

ISBN 978-2-7598-0624-9 EDP Sciences Les UlisISSN 1633-4760

e-ISSN 1638-1963

This work is subject to copyright. All rights are reserved, whether the whole or part of thematerial is concerned, specifically the rights of translation, reprinting, re-use of illustrations,recitation, broad-casting, reproduction on microfilms or in other ways, and storage in data banks.Duplication of this publication or parts thereof is only permitted under the provisions of theFrench Copyright law of March 11, 1957. Violations fall under the prosecution act of the FrenchCopyright Law.

c© EAS, EDP Sciences 2011

Printed in UK

Foreword

Some twenty five years ago, driven by ground-based, airborne, and IRAS ob-servations, the PAH hypothesis was first formulated to explain the strong emis-sion features that dominate the mid-infrared spectra of most bright astronomicalsources. In this hypothesis, the well-known infrared emission features – at 3.3,6.2, 7.7, 8.6, and 11.2 μm – were attributed to large (<50 C-atom) PolycyclicAromatic Hydrocarbon (PAH) molecules that are pumped by the strong far-ultraviolet photon flux from nearby stars. Since then, spectroscopy using theShort Wavelength Spectrometer and ISOCAM on the Infrared Space Observatorylaunched by the European Space Agency in 1995, the InfraRed Spectrograph onthe Spitzer Space Telescope launched by NASA in 2005 and the InfraRed Cameraonboard the Japanese AKARI satellite launched in 2006 have revealed the richnessof the interstellar infrared emission spectrum and the variations therein. Thesespectral variations reflect variations in the molecular characteristics of the PAHfamily, reflecting the local physical and chemical conditions of the emitting re-gions. Parallel to these observational developments, experimental and theoreticalstudies of the physical and chemical properties of astrophysically relevant PAHshave really taken off. Such studies aim at elucidating the intrinsic infrared, visible,and ultraviolet properties of large PAH molecules and their dependence on molec-ular characteristics. In addition, dedicated experiments are performed to studythe origin, evolution, and chemical role of PAHs in the interstellar medium. It isclear that this field has really taken off and the PAH hypothesis has evolved intothe reigning paradigm.

Observations have shown that PAH molecules are abundant and ubiquitous inthe interstellar medium. Conversely, PAHs may also be a dominant “force” in theinterstellar medium, dominating the photoelectric heating of interstellar gas andmay be important for the ionization balance inside dense molecular cloud cores.While much progress has been made, still more remains to be discovered includingthe role of derivative species such as nitrogenated PAHs, PAH clusters and PAHcomplexes with metals, and the relationship between PAH molecules and carbona-ceous grains including Hydrogenated Amorphous Carbon and diamond. PAHs mayalso provide a catalytic surface for the formation of, for example, molecular hy-drogen. Furthermore, photolysis of PAH related species may provide a source ofsmall hydrocarbons particularly in regions rich in UV photons.

We are now reaching a stage where we can start to use the observations of theIR emission spectrum as diagnostic tools to determine the physical conditions inthe emitting regions of, in particular, regions of star and planet formation. Be-cause the IR emission features dominate the IR spectrum of regions of massive starformation, these bands are also often used as proxies to determine the importanceof star formation on galactic scales. Specifically, the importance of star forma-tion versus AGN activity for the luminosity source of (Ultra)Luminous InfraRedGalaxies is based upon a quantitative interpretation of the observed PAH emis-sion from galactic nuclei. At this moment, the Herschel and Planck ESA spacemissions are geared towards probing the low frequency bending modes of PAHs

IV

and the rotational transitions of this spinning dust component. The James WebbSpace Telescope and the Stratospheric Observatory for Infrared Astronomy, andthe possible SPICA mission will provide in the coming years spectral imaging athigh resolution and increased sensitivity. The future for the PAH universe looksbright.

We considered it therefore timely to organize a scientific meeting “PAHs and theuniverse: A Symposium to celebrate the 25 th anniversary of the PAH hypothesis”with the goal to bring together experts in the area of astronomical observations,laboratory studies, and astronomical modeling of interstellar PAHs to discuss thestate-of-the-art and to chart the future. Moreover, and more importantly, twoof the pioneers of the interstellar PAH hypothesis – Lou Allamandola and AlainLeger – are reaching a milestone in their life – they soon will be 65 years young -and this meeting provided a good occasion to celebrate their accomplishments inopening up and driving this field.

The symposium was hosted in Toulouse (France) by the CESR and LCPQ(University of Toulouse and CNRS), May 31 through June 4th, 2010. We arevery grateful for their hospitality and want to acknowledge the hard work bythe local organizing committee that was instrumental in making this sympo-sium such a resounding success. This symposium would not have been possiblewithout the generous financial support by INSU-CNRS, Universite Paul Sabatier(UPS), CNES, Region Midi-Pyrenees, Ministere de l’Enseignement Superieur etde la Recherche, ESA, IRSAMC-UPS, INC-CNRS, Reseau de chimie theorique,INP-CNRS, and Departement de la Haute Garonne. We would like to thank theScientific Organizing Committee for the well-conceived scientific program, whichhighlighted all aspects of interstellar PAHs. We also thank the participants forthe reviews, contributed papers, and poster papers and the stimulating discus-sions. The PAH symposium was attended by some 130 scientists from 18 differentcountries and 5 continents, all united in their quest for the molecular Universe.

This is the first time in 25 years that a symposium was organized on thisimportant topic and we felt that this warranted the publication of proceedings.The goal of the proceedings paralleled those of the symposium: to publish a lastinglegacy of this meeting summarizing the field and charting the future. We haveasked the reviewers to write review chapters at a graduate level pedagogical fashionto yield a reference text for years to come. In addition, these proceedings includethe contributed papers, which provide a cross cut of the field as it is today. Inreading these papers, we have been impressed by the hard work and care withwhich they were written. A symposium and its proceedings are only as good asthe participants and speakers make it and, from that perspective, this meetingwas top-notch.

Mountain View & Toulouse December 9, 2010Xander Tielens & Christine Joblin

DOI: 10.1051/eas/1146000

List of Participants

ACKE Bram Instit. Astronomy, K.U. Leuven, Belgium

ALATA Ivan Centre Laser, Univ. Paris Sud 11, France

ALLAMANDOLA Louis NASA Ames Research Center, CA, USA

ALVARO GALUE Hector FOM Rijnhuizen, The Netherlands

BAOUCHE SAOUD Paris Observatory, France

BARTHEL Robert Instit. Phys. Chem., KIT, Karlsruhe, Germany

BASIRE Marie ISMO, Univ. Paris Sud/CNRS, France

BEARPARK Michael Imperial College London, UK

BECKER Luann Johns Hopkins Univ., Baltimore, USA

BERNARD Jean-Philippe CESR, Univ. Toulouse/CNRS, France

BERNE Olivier Leiden Observatory, The Netherlands

BIENNIER Ludovic IPR, Univ. Rennes 1/CNRS, France

BIERBAUM Veronica Univ. Colorado, Boulder, USA

BLANCHET Valerie LCAR-IRSAMC, Univ. Toulouse/CNRS, France

BOERSMA Christiaan NASA Ames Research Center, CA, USA

BOGGIO-PASQUA Martial LCPQ, Univ. Toulouse/CNRS, France

BOULANGER Francois IAS, Univ. Paris Sud/CNRS, France

BOUWMAN Jordy Leiden Observatory, The Netherlands

BRECHIGNAC Philippe ISMO, Univ. Paris Sud/CNRS, France

BRYSON Kathryn NASA Ames Research Center, CA, USA

CALZETTI Daniela Astronomy Dept., Univ. Massachusetts, USA

CAMI Jan Univ. Western Ontario & SETI Instit., Canada

CANDIAN Alessandra University of Nottingham, UK

CARPENTIER Yvain Max-Planck-Instit. Astron., Jena, Germany

CERNICHARO Jose Centro de Astrobiologia INTA-CSIC, Spain

CHANDRASEKARAN Vijayanand ISM, Univ. Bordeaux 1/CNRS, France

CHERCHNEFF Isabelle Universitaet Basel, Switzerland

CONTRERAS Cesar NASA Ames Research Center, CA, USA

COUPEAUD Anne CESR, Univ. Toulouse/CNRS, France

COX Nick Instit. Astronomy, K.U. Leuven, Belgium

D’HENDECOURT Louis IAS, Univ. Paris Sud/CNRS, France

DALLE ORE Cristina SETI Institute/ NASA Ames, CA, USA

DARTOIS Emmanuel IAS, Univ. Paris Sud/CNRS, France

DEDONDER Claude CNRS/Univ. Paris-Sud 11, France

DEMYK Karine CESR, Univ. Toulouse/CNRS, France

DOPFER Otto TU Berlin, Germany

DRAINE Bruce Princeton University, USA

ELLINGER Yves LCT - Univ. Pierre et Marie Curie/ CNRS, France

FERAUD Geraldine ISMO, Univ. Paris Sud/CNRS, France

FRIHA Hela ISMO, Univ. Paris Sud/CNRS, France

GADALLAH Kamel University of Jena, Germany

GADEA Florent Xavier LCPQ, Univ. Toulouse/CNRS, France

GADALLAH Kamel University of Jena, Germany

VI

GADEA Florent Xavier LCPQ, Univ. Toulouse/CNRS, France

GALLIANO Frederic AIM, CEA/Saclay, France

GEBALLE Tom Gemini Observatory, Hawaii, USA

GEPPERT Wolf Stockholm University, Sweden

GIARD Martin CESR, Univ. Toulouse/CNRS, France

GODARD Marie IAS, Univ. Paris Sud/CNRS, France

GORDON Karl Space Telescope Science Instit., Baltimore, USA

GUENNOUN Zohra ISM, Univ. Bordeaux 1/CNRS, France

HAMMONDS Mark University of Nottingham, UK

HANNACHI Yacine ISM, Univ. Bordeaux 1/CNRS, France

HARAGUCHI Kentaro Nagoya University, Japan

HEBDEN Kerry University of Manchester, UK

HELTON L. Andrew University of Minnesota, Minneapolis, USA

HERBST Eric Ohio State University, Columbus, USA

HOLM Anne I. S. Stockholm University, Sweden

HONY Sacha AIM, CEA/Saclay, France

HORNEKAER Liv Aarhus University, Denmark

HUNT Leslie INAF-Osserv. Astrof. Arcetri, Italy

IGLESIAS-GROTH Susana Instit. Astrof¡sica de Canarias, Spain

IVANOVSKAYA Viktoria ISMO, Univ. Paris Sud/CNRS, France

JAGER Cornelia Friedrich Schiller Univ., Jena, Germany

JOALLAND Baptiste CESR/LCPQ - Univ. Toulouse/CNRS, France

JOBLIN Christine CESR, Univ. Toulouse/CNRS, France

JØRGENSEN Bjarke Aarhus University, Denmark

JOUVET Christophe ISMO, Univ. Paris Sud/CNRS, France

KAMP Inga Kapteyn Astron. Instit., Groningen, The Netherlands

KANEDA Hidehiro Nagoya University, Japan

KAZMIERCZAK Maja Nicolaus Copernicus Univ., Torun, Poland

KIM Ji Hoon Seoul National Univ., Republic of Korea

KLÆRKE Benedikte Aarhus University, Denmark

KRELOWSKI Jacek Nicolaus Copernicus Univ., Torun, Poland

LAGACHE Guilaine IAS, Univ. Paris Sud/CNRS, France

LAWICKI Arkadiusz CIMAP - GANIL, France

LE BOURLOT Jacques Univ. Paris-Diderot, Paris 7 & Obs. Paris, France

LE PADELLEC Arnaud CESR, Univ. Toulouse/CNRS, France

LE PAGE Valery Univ. Colorado, Boulder, USA

LEE Timothy NASA Ames Research Center, CA, USA

LEGER Alain IAS, Univ. Paris Sud/CNRS, France

LI Aigen University of Missouri, Columbia, USA

MAHADEVAPPA Naganathappa Swami Ramanand Teerth Marathwada Univ., India

MARBLE Andy Steward Observatory, Tucson, USA

MASCETTI Joelle ISM, Univ. Bordeaux 1/CNRS, France

MASON Rachel Gemini Observatory, Hawaii, USA

MAURETTE Michel CSNSM, Univ. Paris Sud/CNRS, France

MAYER Paul Chemistry Dep., Univ. Ottawa, Canada

MENNELLA Vito INAF-Osserv. Astron. Capodimonte, Italy

MISSELT Karl Steward Observatory, Tucson, USA

VII

MONTILLAUD Julien CESR, Univ. Toulouse/CNRS, France

MULAS Giacomo INAF - Osserv. Astron. Cagliari, Italy

NILSSON Louis Aarhus University, Denmark

O’DOWD Matt Columbia University, New York, USA

OHSAWA Ryou Grad. School of Science, Univ. Tokyo, Japan

OKADA Yoko I. Physik. Institut, Univ. Koln, Germany

OMONT Alain IAP, Univ. Pierre et Marie Curie/CNRS, France

ONAKA Takashi University of Tokyo, Japan

OOMENS Jos FOM Rijnhuizen, The Netherlands

OTAGURO Jacqueline Western Ontario Univ., London, Canada

PARNEIX Pascal ISMO, Univ. Paris Sud/CNRS, France

PATHAK Amit Indian Instit. Astrophysics, Bangalore, India

PAUZAT Francoise LCT - Univ. Pierre et Marie Curie/ CNRS, France

PEETERS Els Univ. Western Ontario & SETI Instit., Canada

PILLERI Paolo CESR, Univ. Toulouse/CNRS, France

PINO Thomas ISMO, Univ. Paris Sud/CNRS, France

PUGET Jean-Loup IAS, Univ. Paris Sud/CNRS, France

RAPACIOLI Mathias LCPQ, Univ. Toulouse/CNRS, France

RASTOGI Shantanu Physics Dept., DDU Gorakhpur Univ., India

RHO Jeonghee California Institute of Technology, USA

RICKETTS Claire NASA Ames Research Center, CA, USA

RISTORCELLI Isabelle CESR, Univ. Toulouse/CNRS, France

ROUILLE Gael Universitat Jena, Germany

ROUSSEAU Patrick CIMAP - GANIL, France

SAKON Itsuki University of Tokyo, Japan

SALAMA Farid NASA Ames Research Center, CA, USA

SALES Dinalva Aires Univer. Federal Rio Grande do Sul, Porto Alegre, Brazil

SANDSTROM Karin Max-Planck-Instit. Astron., Heidelberg, Germany

SARRE Peter University of Nottingham, UK

SELLGREN Kris Ohio State University, Columbus, USA

SHENOY Sachindev NASA Ames Research Center, CA, USA

SIEBENMORGEN Ralf ESO, Garching. Germany

SIMON Aude LCPQ, Univ. Toulouse/CNRS, France

SMITH Erin NASA Ames Research Center, CA, USA

SMOLDERS Kristof Instit. Astronomy, K.U. Leuven, Belgium

SNOW Theodore Univ. Colorado, Boulder, USA

SPIEGELMAN Fernand LCPQ, Univ. Toulouse/CNRS, France

STEGLICH Mathias Universitat Jena, Germany

TALBI Dahbia GRAAL, Univ. Montpellier 2/CNRS, France

TEILLET-BILLY Dominique ISMO, Univ. Paris Sud/CNRS, France

THROWER John Aarhus University, Denmark

TIELENS Alexander Leiden Observatory, The Netherlands

VERSTRAETE Laurent IAS, Univ. Paris Sud/CNRS, France

WALKER Mark Manly Astrophysics, Australia

WOODS Paul University of Manchester, UK

Contents

Foreword . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . III

List of participants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . V

Introduction

25 Years of PAH HypothesisA.G.G.M. Tielens. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . 3

Rich IR Spectra of Interstellar PAHs

Astronomical Observations of the PAH Emission BandsE. Peeters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

Astronomical Models of PAHs and DustB.T. Draine . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29

Dialectics of the PAH Abundance Trend with MetallicityF. Galliano . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43

The Shape of Mid-IR PAH Bands in the UniverseO. Berne, P. Pilleri and C. Joblin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49

AKARI Near-Infrared Spectroscopy of 3 Micron PAH and 4 MicronPAD Features

T. Onaka, I. Sakon, R. Ohsawa, T. Shimonishi, Y. Okada,M. Tanaka and H. Kaneda. . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . 55

Laboratory Infrared Spectroscopy of PAHsJ. Oomens. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61

Computational IR Spectroscopy for PAHs: From the Early Yearsto the Present Status

F. Pauzat . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75

X

Modeling the Anharmonic Infrared Emission Spectra of PAHs: Applicationto the Pyrene Cation

M. Basire, P. Parneix, T. Pino, Ph. Brechignac and F. Calvo . . . . . . . . 95

Laboratory Spectroscopy of Protonated PAH Molecules RelevantFor Interstellar Chemistry

O. Dopfer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103

The NASA Ames PAH IR Spectroscopic Database and the far-IRC. Boersma, L.J. Allamandola, C.W. Bauschlicher, Jr., A. Ricca,J. Cami, E. Peeters, F. Sanchez de Armas, G. Puerta Saborido,A.L. Mattioda and D.M. Hudgins. . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . 109

Analyzing Astronomical Observations with the NASA Ames PAHDatabase

J. Cami . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117

Search for far-IR PAH Bands with Herschel: Modellingand Observational Approaches

C. Joblin, G. Mulas, G. Malloci and E. Bergin & the HEXOSconsortium . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123

PAHs and Star Formation in the Near and Far Universe

Polycyclic Aromatic Hydrocarbons as Star Formation Rate IndicatorsD. Calzetti . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133

PAHs and the ISM in Metal-Poor StarburstsL.K. Hunt, Y.I. Izotov, M. Sauvage and T.X. Thuan . . . . . . . . . . . . . . . . 143

Introduction to AMUSES: AKARI Survey with a Window of OpportunityJ.H. Kim, M. Im, H.M. Lee, M.G. Lee and the AMUSES team . . . . . . 149

The Lifecycle of PAHs in Space

PAH Evolution in the Harsh Environment of the ISMH. Kaneda, T. Onaka, I. Sakon, D. Ishihara, A. Mouri,M. Yamagishi and A. Yasuda . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 157

PAH and Dust Processing in Supernova RemnantsJ. Rho, M. Andersen, A. Tappe, W.T. Reach, J.P. Bernardand J. Hewitt . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169

XI

The Formation of Polycyclic Aromatic Hydrocarbons in EvolvedCircumstellar Environments

I. Cherchneff . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . 177

Insights into the Condensation of PAHs in the Envelope of IRC +10216L. Biennier, H. Sabbah, S.J. Klippenstein, V. Chandrasekaran,I.R. Sims and B.R. Rowe . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . 191

Formation and Evolution of Circumstellar and Interstellar PAHs:A Laboratory Study

C.S. Contreras, C.L. Ricketts and F. Salama . . . . . . . . . . . . . . . . . . . . . . . . 201

Confirmation of C60 in the Reflection Nebula NGC 7023K. Sellgren, M.W. Werner, J.G. Ingalls, J.D.T. Smith,T.M. Carleton and C. Joblin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 209

The Spitzer Surveys of the Small Magellanic Cloud: Insights intothe Life-Cycle of Polycyclic Aromatic Hydrocarbons

K.M. Sandstrom, A.D. Bolatto, B.T. Draine, C. Botand S. Stanimirovic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 215

PAH-related Very Small Grains in Photodissociation Regions: Implicationsfrom Molecular Simulations

M. Rapacioli, F. Spiegelman, B. Joalland, A. Simon, A. Mirtschink,C. Joblin, J. Montillaud, O. Berne and D. Talbi . . . . . . . . . . . . . . . . . . . . 223

The Formation of Benzene in Dense EnvironmentsP.M. Woods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 235

Experimental Studies of the Dissociative Recombination Processesfor the C6D+

6 and C6D+7 Ions

M. Hamberg, E. Vigren, R.D. Thomas, V. Zhaunerchyk, M. Zhang,S. Trippel, M. Kaminska, I. Kashperka, M. af Ugglas, A. Kallberg,A. Simonsson, A. Paal, J. Semaniak, M. Larsson and W.D. Geppert 241

VUV Photochemistry of PAHs Trapped in Interstellar Water IceJ. Bouwman, H.M. Cuppen, L.J. Allamandola and H. Linnartz . . .. . . 251

PAHs in Regions of Planet Formation

Observations of Hydrocarbon Emission in Disks Around Young StarsB. Acke . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 259

Evolution of PAHs in Protoplanetary DisksI. Kamp . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 271

XII

PAH in Vectorized Three Dimensional Monte Carlo Dust RadiativeTransfer Models

R. Siebenmorgen, F. Heymann and E. Krugel. . . . . . . . . . . .. . . . . . . . . . . . 285

PAHs and Carbonaceous Grains & Solar System Materials

From PAHs to Solid CarbonC. Jager, H. Mutschke, T. Henning and F. Huisken . . . . . . . . .. . . . . . . . . 293

PAHs and AstrobiologyL.J. Allamandola . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . 305

Solid State Molecular Reactors in SpaceM. Maurette . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 319

Polycyclic Aromatic Hydrocarbons and the Extinction CurveG. Mulas, G. Malloci, C. Joblin and C. Cecchi–Pestellini . . . . . . . . . . . . 327

The Diffuse Interstellar Bands in History and in the UVT.P. Snow and J.D. Destree . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 341

The PAH-DIB HypothesisN.L.J. Cox . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 349

Electronic Spectroscopy of PAHsT. Pino, Y. Carpentier, G. Feraud, H. Friha, D.L. Kokkin,T.P. Troy, N. Chalyavi, Ph. Brechignac and T.W. Schmidt . . . . .. . . . . 355

Spectroscopy of Protonated and Deprotonated PAHsM. Hammonds, A. Pathak, A. Candian and P.J. Sarre . . . . . . . . . . . . . . 373

Observations of Interstellar Carbon CompoundsE. Dartois . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 381

Interaction of Atomic Hydrogen with Carbon GrainsV. Mennella . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 393

Near-Infrared Spectroscopy of Interstellar DustF. Boulanger, T. Onaka, P. Pilleri and C. Joblin . . . . . . . . . . . . . . . . . . . . 399

Atypical Dust Species in the Ejecta of Classical NovaeL.A. Helton, A. Evans, C.E. Woodward and R.D. Gehrz . . . . . . . . . . . . 407

XIII

The Role of PAHs in the Interstellar Medium

The Role of PAHs in the Physics of the Interstellar MediumL. Verstraete . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . 415

PAHs and the Chemistry of the ISMV.M. Bierbaum, V. Le Page and T.P. Snow. . . . . . . . . . . . .. . . . . . . . . . . . . 427

[FePAH]+ Complexes and [FexPAHy]+ Clusters in the Interstellar Medium:Stability and Spectroscopy

A. Simon, M. Rapacioli, F. Spiegelman and C. Joblin . . . . . . . .. . . . . . . . 441

Modelling the Physical and Chemical Evolution of PAHs and PAH-relatedSpecies in Astrophysical Environments

J. Montillaud, C. Joblin and D. Toublanc. . . . . . . . . . . . . .. . . . . . . . . . . . . . 447

Superhydrogenated PAHs: Catalytic Formation of H2

J.D. Thrower, L. Nilsson, B. Jørgensen, S. Baouche, R. Balog,A.C. Luntz, I. Stensgaard, E. Rauls and L. Hornekær . . . . . . . .. . . . . . . . 453

Summary of the Meeting

Summary of the MeetingA. Omont . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 463

Author Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 469

Astronomical Object Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 471

Chemical Compound Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 473

Subject Index . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . 475

Introduction

PAHs and the UniverseC. Joblin and A.G.G.M. Tielens (eds)EAS Publications Series, 46 (2011) 3-10

25 YEARS OF PAH HYPOTHESIS

A.G.G.M. Tielens1

Abstract. Strong IR emission features at 3.3, 6.2, 7.7, 8.6, and 11.2 μmare a common characteristic of the interstellar medium of the MilkyWay and nearby galaxies and out to redshifts of 3. Here, we reviewthe history of the PAH hypothesis which attributes these emission fea-tures to vibrational fluorescence of large (∼ 50 carbon-atom PolycyclicAromatic Hydrocarbon molecules pumped by ultraviolet photons fromnearby stars or the average interstellar radiation field. Over the last25 years, our insight in the characteristics of these molecules and theirrole in the Universe has greatly improved and the PAH hypothesis isalive and well; not in the least due a remarkable adaptability. Notsurprisingly, the precise characteristics of these species remains to bedefined.

1 The early years

Over the last three decades our understanding of the interstellar medium has in-creased dramatically. To a large extent this has been driven by rapid develop-ments in infrared detector technology driven by the Cold War. Sensitive infraredspectrometers operating in all the infrared windows are now standard on all majorground-based infrared observatories and major infrared space missions have probedthe Universe at all infrared wavelengths. As soon as the infrared sky opened upfor spectroscopy in the early 70ies, Gillett et al. (1973) discovered broad emis-sion features in the L (3 μm) and N (10 μm) band windows (Fig. 1). While agood fit could be obtained to the profile and peak position of the 11.3 μm bandwith carbonates, air-borne observatories refuted this identification since the strongband at 6.85 μm was not detected (Russell et al. 1977; Willner et al. 1979). In-stead, these observations revealed that the the 3.3 and 11.3 μm bands were partof a set of strong bands – together with bands at 6.2, 7.7, and 8.6 μm – that oc-cured together. These and subsequent observations established that these broad

1 Leiden Observatory, Leiden University, PO Box 9513, Leiden, The Netherlands

c© EAS, EDP Sciences 2011DOI: 10.1051/eas/1146001

4 PAHs and the Universe

Fig. 1. The first overview spectrum of the PAH emission bands obtained from the ground

in the L and N band windows and from the Kuiper Airborne Observatory (5–8 μm).

Taken from the review by Willner et al. (1979).

IR emission features were a common characteristic of all objects where strong ul-traviolet (UV) photons illuminated nearby material. Surprisingly, the nature ofthe objects seemed irrelevant: planetary nebulae, HII regions, reflection nebulae,and starburst galaxies all showed the same set of emission features. For almost adecade, identification of the carrier remained mysterious and these bands were col-lectively called the Unidentified InfraRed (UIR) bands. In the early 80ies, Duley& Williams (1981) realized that the peak positions of the UIR bands were verycharacteristic for aromatic materials and they proposed Hydrogenated AmorphousCarbon grains as the carriers. This proposal highlighted, however, an energeticissue inherent to all “bulk” grain approaches for the UIR bands: The emissionoriginated far from the illuminating star where the radiative equilibrium tempera-ture of grains is much too low (Td ∼ 10−75 K) to account for emission in the near-and mid-IR range. The next – and key step – was provided by Kris Sellgren 1984who studied the IR emission from reflection nebula where this “energy” crisis wasparticularly rampant. She realized that the limited heat capacity of very smallgrains would result in large temperature excursions upon absorption of a singleUV photon. Assuming a typical UV photon energy of 10 eV, a size of 50 C-atomswas implied. Independently, Leger & Puget (1984) and Allamandola et al. (1985)put these concepts together and the PAH hypothesis was born: the IR emissionfeatures at 3.3, 6.2, 7.7, and 11.3 μm are carried by large Polycyclic AromaticHydrocarbon molecules containing some 50 C-atoms and with a drived abundance

A.G.G.M. Tielens: PAH Hypothesis 5

of 5–10% of the elemental carbon, these molecules – as a class – are the mostabundant organic molecules in space.

In these early years, there was much discussion on the character of the carriersof the UIR bands and a plentitude of materials were proposed including PAHmolecules, HAC grains, Quenched Carbonaceous Composite, and coal. At theSanta Clara conference at the end of the 80ies, the situation was summarized bythe title of the session on the IR emission bands: “the Overidentified InfraRedbands” (Allamandola & Tielens 1989). All of these materials shared howeverthe basic molecular structure of PAH molecules: benzene rings fused togetherinto aromatic structures decorated with hydrogen at the edge, although many ofthese materials also contained other molecular structures often with other elementsbesides C and H. While the idea that the IR emission bands are carried by largemolecules rather than grains is now undisputed, the exact structure of the speciesand in particular the presence of heteroatoms are still widely and vividly debated.

2 The PAH hypothesis

This heralded an exciting period were new ideas were announced at a rapid pace.In particular, at about the same time Kris Sellgren performed her breakthroughstudy, the IRAS mission discovered widespread emission at 12 and 25 μm in thediffuse interstellar medium of the Milky Way and other galaxies (Low et al. 1984).While there was no spectrum, it was quickly surmised that this emission was, inessence, the UIR bands at 7.7, 8.6, and 11.3 μm in combination with a continuumsteeply rising towards longer wavelengths and all pumped by the interstellar ra-diation field (Puget et al. 1985). Hence, PAH molecules had to be ubiquitousthroughout the interstellar medium. It was also realized that such moleculeswould also have absorption bands in the visible and UV range. In fact, the well-known Diffuse Interstellar Bands (DIBs) in the visible region of the spectrum mightwell be due to ionized or “radical” members of the PAH family (van der Zwet &Allamandola 1985; Crawford et al. 1985; Leger et al. 1985) and these bands maywell hold important clues for the identification of individual interstellar PAHs. Be-cause of their high abundance, it was also quickly clear that PAH molecules wouldhave an enormous influence on the energy and ionization balance of interstellargas both in the diffuse and dense media and hence on the phase structure, gasphase composition, and ambipolar diffusion of clouds (Verstraete et al. 1990; Lepp& Dalgarno 1988). Many of these ideas are summarized in the first reviews (Leger& Puget 1989; Allamandola et al. 1989). At this point, the PAH hypothesis canbe summarized as:

• A substantial fraction of the carbon in the universe is in form of PAHmolecules.

• A substantial fraction of the UV-visible radiative stellar energy in galaxiesis “transformed” by PAHs to the mid-IR range.


Fig. 2. The mid-infrared spectra of the PhotoDissociation Region in the Orion Bar and

in the Planetary Nebulae, NGC 7027 are dominated by a rich set of emission features.

Assignments of these features with vibrational modes of PAH molecules are labeled at

the top. Figure adapted from Peeters et al. (2002).

• PAH molecules may provide observational signatures throughout the fullspectral range from the extreme ultraviolet to the radio regime.

• PAHs are destroyed by hard UV radiation and thus they must be at thebasis of a complex chemistry in the ISM.

• PAHs contribute to many key physical processes that affect the basic struc-ture and evolution of the interstellar medium of galaxies.

3 The space-based era

The opening of the full infrared window by spectroscopic space missions further in-creased our knowledge of the interstellar PAHs. The Low Resolution Spectrometer(LRS) – a slitless spectrometer sensitive from 7.5 to 23 μm with a resolving powerof R ∼ 20– on the InfraRed Astronomy Satellite (IRAS) has provided a firstsystematic overview (Olnon 1986) of the extent of the UIR bands. The ShortWavelength Spectrometer, SWS (de Graauw et al. 1996) on board of the Infrared


Space Observatory (ISO), represented the next big step forward by providing com-plete 2.5 to 45 μm spectra of essentially all IR-luminous galactic sources withresolutions ranging from ∼200 up to ∼2000 (for the brightest sources). The ISOmission also carried the ISOCAM (Cesarsky et al. 1996) aboard – which in CVF-mode provided spectral-imaging capabilities over ∼1 square arcminute areas withpixel sizes of 3–6 arcseconds at a modest spectral resolution of 50 – that pro-duced equally important studies of interstellar PAHs. Together these instrumentsrevealed the dominance of the UIR bands for essentially all bright sources thatthey could study. The InfraRed Spectrometer (IRS) on board of the Spitzer SpaceTelescope – operating from 5.2 to 38 μm at low spectral resolution (60 to 130)and from 10 to 37 μm at moderate resolution (R ∼ 600) but with greatly supe-rior sensitivity over ISO (Houck et al. 2004) – allowed systematic spectroscopicstudies of large samples of even weak sources in the Milky Way, nearby galaxiesand in bright galactic nuclei out to redshifts of ∼3. All of these studies togetherdemonstrated that PAHs dominate the spectra of almost all objects, includingHII regions, reflection nebulae, young stellar objects, Planetary Nebulae, post-asymptotic-giant-branch objects, nuclei of galaxies, and ultraluminous infraredgalaxies (ULIRGs). Moreover, the IR cirrus, the surfaces of dark clouds, and thegeneral interstellar medium (ISM) of galaxies are set aglow in these IR emissionfeatures (for a recent review see Tielens 2005).

4 PAHs in the far Universe

The high sensitivity of the Spitzer instruments has demonstrated that PAH emis-sion is present out to redshifts of z � 3. Regular and starburst galaxies in the nearUniverse are readily surveyed with Spitzer and for example the SINGS sample aswell as open and guaranteed time programs are rife with galaxies whose spectraare dominated by the mid-IR emission features (e.g., Smith et al. 2007). Withthese studies, reliable criteria could be established that allow the use of the PAHemission strength as tracers of the star formation rate (see Calzetti, elsewhere inthis volume). With IRS/Spitzer, galaxies much farther away could still be studiedin the case of lens systems. The cloverleaf system at a redshift of 2.58 is a primeexample (Fig. 3; Lutz et al. 2007) but many more studies exist (Rigby et al. 2008;Lutz et al. 2005). As a class, the ultraluminous sub-millimeter galaxies were alsowithin reach out to redshifts of ∼3 with Spitzer (Pope et al. 2008).

In a more indirect way, Spitzer probed distant but less luminous systemsthrough source counts. The spectrum of a galaxy will shift further into the in-frared as its distance increases. Because the PAH features dominate the mid5–15 μm region and the continuum due to very small grains does not start to riseuntil 30 μm, there is a “trough” in the emission from normal or starburst galax-ies (e.g., non-ULIRGS). Then, as the PAH features shift into the 24 μm MIPSband at a redshift of z � 2, this will lead to a “bump” in the source counts withflux (e.g., distance; Fig. 4; Lagache et al. 2004). These studies demonstrate thatPAHs are present at abundances of �10% of the elemental C in the star forminggalaxies as far as Spitzer could probe and that the typical galaxy in these studies


Fig. 3. The infrared spectrum of the Cloverleaf system is characterized by a strong

continuum – reminiscent of AGN sources – with distinct PAH features superimposed.

The strength of these PAH features indicates a star formation rate of some 103 M�/yr

(Lutz et al. 2007).

Fig. 4. (a) The K-band correction in different filters shows a “bump” when the PAH

emission features shift through the wavelength range covered. (b) Redshift contribution

to the number counts at 24 μm. The dotted, dashed, dash-dotted, tripledot-dashed, and

long-dashed lines correspond to the number counts up to redshifts 0.3, 0.8, 1, 1.3, and 2,

respectively. Taken from Lagache et al. (2006) & (2004).

(L ∼ 1011 − 1012 L�; star formation rates 20− 130 M�/yr); M � 1010− 1011 M�has a mid-IR spectrum dominated by the PAH features. Moreover, the CosmicInfraRed Background between 24 and 160 μm is dominated by these galaxies.


5 The future

With Herschel a new era has arrived and the study of the Universe at far-infraredand sub-millimeter wavelengths has come within reach. First and foremost,Herschel will provide a deeper understanding of the dark and cold or lukewarmlocal Universe through spectroscopic studies of e.g., [OI], [CII], and CO lines innearby galactic and extragalactic PAH-bright regions. This will link the mid-IRPAH characteristics in a direct way to the local physical conditions. In addition,spectroscopic studies may reveal far-IR spectral signatures of interstellar PAHswhich might provide key information on the individual PAH molecules present inspace (see Joblin et al., elsewhere in this volume). Finally, with Herschel, deepsource counts at long wavelengths, may reveal the signatures of PAH emissionbands as they shift throught the 70 μm PACS band at a redshift of ∼7. Indeed, atthis point it is fair to expand the PAH hypothesis to its strongest version to read:

• PAH molecules formed in the ejecta of the first generation of stars pollutingthe Universe with heavy atoms in the form of molecules and dust.

• Ever since that time, PAHs are a key component of the “dust” budget ofgalaxies.

• Ever since that time, PAH molecules have converted a substantial fractionof the UV-visible radiative stellar energy in galaxies to the mid-IR range.

• Ever since that time, PAH molecules are at the basis of a complex chemistryin the ISM of galaxies.

• Ever since that time, PAHs contribute to many key physical processes thataffect the basic structure and evolution of the interstellar medium of galaxies.

In short, PAHs are abundant and important throughout the full star forming eraof the Universe.

6 Summary

Over the last 25 years, the PAH hypothesis has blossomed thanks to a multitudeof observational, experimental, and theoretical studies and the dedicated efforts ofmany scientists. This conference overviews the current status of the field throughthe many reviews of key topics related to interstellar PAHs and a cross cut of theactive research taking place through the many contributed papers. Overall, thePAH hypothesis has shown to be very robust (or adaptable for the more cynicallyinclined) and “weathered” the many new observational discoveries and insightsthat have developed over time. This of course, merely presents a challenge tothe younger generation to expand on it so much that the originators might notrecognize it anymore or even completely disprove it!


References

Allamandola, L.J., & Tielens, A.G.G.M., Interstellar Dust (Kluwer: Dordrecht)

Allamandola, L.J., Tielens, A.G.G.M., & Barker, J.R., 1985, ApJ, 290, L25

Allamandola, L.J., Tielens, A.G.G.M., & Barker, J., 1989, ApJS, 71, 733

Cesarsky, C., Abergel, A., Agnese, P., et al., 1996, A&A, 315, L32

Crawford, Tielens, A.G.G.M., & Allamandola, L.J., 1985, ApJ, 293, L45

Duley, W., & Williams, D.A., 1981, MNRAS, 196, 269

Gillett, F.C., Forest, W.J., & Merrill, K.M., 1973, ApJ, 183, 87

de Graauw, Th., Haser, L.N., Beintema, D.A., et al., 1996, A&A, 315, L49

Houck, J., Roellig, T.L., van Cleve, J., et al., 2004, ApJS, 154, 18

Lagache, G., Dole, H., Puget, J.-L., et al., 2004, ApJS, 154, 112

Lagache, G., Puget, J.-L., & Dole, H, 2006, ARAA, 43, 727

Leger, A., & d’Hendecourt, L., 1985, A&A, 146, 81

Leger, A., & Puget, J.-L., 1984, A&A, 137, L5

Lepp, S., & Dalgarno, A., 1988, ApJ, 324, 553

Low, F.J., Young, E., Beintema, D.A., et al., 1984, ApJ, 278, L19

Lutz, D., Yan, L., Armus, L., et al., 2005, ApJ, 632, L13

Lutz, D., Sturm, E., Tacconi, L.J., et al., 2007, ApJ, 661, L25

Olnon, F., Raimond, E., Neugebauer, G., et al., 1986, A&AS, 65, 607

Peeters, E., Hony, S., Van Kerckhoven, C., et al., 2002, A&A, 390, 1089

Pope, A., Chary, R.-R., Alexander, D.M., et al., 2008, ApJ, 675, 1171

Puget, J.-L., Leger, A., & Boulanger, F., 1985, A&A, 142, L19

Puget, J.-L., & Leger, A., 1989, ARAA, 27, 161

Rigby, J.R., Marcillac, D., Egami, E., et al., 2008, ApJ, 675, 262

Russel, R., Soifer, B.T., & Willner, S.P., 1977, ApJ, 217, 149

Sellgren, K., 1984, ApJ, 277, 623

Smith, J.D.T., Draine, B.T., Dale, D.A., et al., 2007, ApJ, 656, 770

Tielens, A.G.G.M., 2008, ARAA, 46, 289

van der Zwet, G., & Allamandola, L.J., 1985, A&A, 146, 76

Verstraete, L., Leger, A., D’Hendecourt, L., Defourneau, D., & Dutuit, O., 1990, A&A,237, 436

Willner, S.P., Puetter, R.C., Russell, R.W., Soifer, B.T., 1979, ApSS, 65, 95

Rich IR Spectra

of Interstellar PAHs


ASTRONOMICAL OBSERVATIONS OF THE PAH EMISSIONBANDS

E. Peeters1, 2

Abstract. The infrared (IR) spectra of many galactic and extragalac-tic objects are dominated by emission features at 3.3, 6.2, 7.7, 8.6and 11.2 μm, generally attributed to the IR fluorescence of PolycyclicAromatic Hydrocarbon molecules (PAHs). These PAH bands havebeen found in a wide variety of environments throughout theUniverse and contain up to 10% of the total power output of star-forming galaxies.

Ground-based telescopes, the Infrared Space Observatory (ISO)and the Spitzer Space Telescope revealed a plethora of weaker PAHbands and provided extensive evidence for significant variability in thePAH spectrum from source to source and spatially within sources. Herewe review the spectral characteristics of astronomical PAHs, their de-pendence on the local physical conditions and the implications for thephysical and chemical characteristics of the carriers.

1 Introduction

The IR spectra of objects associated with dust and gas – including evolved stars,reflection nebulae (RN), the interstellar medium (ISM), star-forming regions, andgalaxies – are dominated by emission features at 3.3, 6.2, 7.7, 8.6 and 11.2 μm, theso-called unidentified infrared (UIR) bands. These bands are generally attributedto the IR fluorescence of Polycyclic Aromatic Hydrocarbon molecules (PAHs) andrelated species3. Observations show without any doubt that PAHs pervade theUniverse. A remarkable 20–30% of the galactic IR radiation (and hence ∼10–15%of the total galactic power budget) is emitted in the PAH bands and 10–20% of thecosmic carbon is locked up in these molecules. Hence, it may not come as a surprise

1 Department of Physics and Astronomy, University of Western Ontario, London, ON N6A3K7, Canada2 SETI Institute, 189 N. Bernardo Ave., Suite 100, Mountain View, CA 94043, USA

3In this paper, the term PAH bands/features is used to refer to these IR emission features.



that PAHs play a crucial role in several astrophysical and astrochemical processesas, for example, heating and cooling of the ISM and grain surface chemistry. Inaddition, PAH intensities are used to determine the star formation rate in galaxies,one of the key indicators for understanding galaxy formation and evolution (cf.Calzetti elsewhere in this volume).

Although the presence of PAHs in space is now generally accepted and PAHsare recognized as critical players in various processes, the characteristics of theemitting population remain unclear. Indeed, no single PAH molecule has beenfirmly identified to date. This identification is hampered by the fact that theirmain spectral signatures, the UIR bands, are due to vibrational modes primarilygoverned by the local physical conditions of the molecular environment (which isthe same for all PAHs), and only show weak dependence on size and structure.However, there is no reason to despair. Although, to first order, the PAH bandsare remarkably similar throughout the Universe, detailed observational studieshave revealed a multitude of spectral variations in peak positions, shapes and(relative) intensities. This provides a tool to determine or further constrain thecharacteristics of the emitting population and to probe the physical conditionsin the emitting regions. When properly calibrated, the PAH emission bands canserve as a probe of the physical conditions in regions near and far. Given theiromnipresence, a PAH diagnostic toolbox will be extremely powerful.

Here, we give a concise overview of the observational properties of the PAHemission bands. Section 2 describes the astronomical PAH spectrum and Section 3highlights the observed spectral characteristics of the PAH bands in terms of theirintensity, peak position and profile. Section 4 reports on the spectral decomposi-tion of the PAH spectra. The astronomical implications are discussed in Section 5and Section 6 reviews the application of the PAH bands as a diagnostic tool. Weconclude with a short summary and a look into the bright PAH future in Section 7.

2 The rich PAH spectrum

In the early 1970’s, new IR instruments on ground-based and airborne telescopesrevealed broad emission features at 3.3, 6.2, 7.7, 8.6 and 11.2 μm (e.g. Gillettet al. 1973; Geballe et al. 1985). In the past two decades, the unprecedentedviews of the IR Universe offered by ground-based telescopes, the Infrared SpaceObservatory (ISO) and the Spitzer Space Telescope showcased the spectral wealthof the PAH spectrum (see Fig. 1). Besides the main PAH bands, a plethora ofweaker bands4 may be present. In addition, deuterated PAHs, PADs, have beententatively detected at 4.4 and 4.65 μm (Peeters et al. 2004a). These bandsare characteristic for the vibrational modes of PAHs (Fig. 1). In particular, the

4This can include bands at 3.4, 3.5, 5.25, 5.75, 6.0, 6.6, 6.9, 7.2-7.4, 8.2, 10.5, 10.8, 11.0, 12.0,13.5, 14.2, 15.8, 16.4, 16.6, 17.0, 17.4 and, 17.8 μm. A band at 19 μm was originally considered aspart of the PAH spectrum. However, its spatial distribution in the reflection nebula NGC 7023 issignificantly distinct from that of the PAH bands (Sellgren et al. 2007). This feature is recentlyfirmly assigned to C60 (Cami et al. 2010; Sellgren et al. 2010, elsewhere in this volume).

E. Peeters: Observations of PAH Bands 15

Fig. 1. The ISO-SWS spectra of the planetary nebula NGC 7027 and the Photo-

Dissociation Region (PDR) at the Orion Bar and, the Spitzer-IRS spectrum of the nuclear

region of NGC 4536, an H ii -type galaxy, (SINGS legacy program; Kennicutt et al. 2003)

illustrate the richness and variety of the PAH spectrum. Also indicated are the aromatic

mode identifications of the major PAH bands.

3.3 μm band is due to CH stretching modes and the 6.2 μm band to CC stretchingmodes. C-H in plane bending modes are responsible for the 8.6 μm band, coupledCC stretching and CH in-plane bending modes for the 7.7 μm complex and the CHout-of-plane bending modes for emission in the 10–15 μm region. At wavelengthslongward of 15 μm, the vibrations involve motions of the entire molecular skeletonand hence are more specific to each molecule. Several PAH bands exhibit highlyasymmetric band profiles with a steep blue rise and a red-shaded wing (e.g., 3.4,5.25, 6.2, 11.2 μm; Barker et al. 1987; Roche et al. 1996; Pech et al. 2002;Verstraete et al. 2001) while other PAH bands exhibit fairly symmetric profiles(e.g., 3.3, 8.6 μm bands). The asymmetry of the 12.7 μm band is exceptionalas it shows a blue-shaded wing with a steep red decline. The 7.7 μm complex isunique and is comprised of at least 2 components: the first peaking at ∼7.6 μmand the second between 7.8–8 μm (e.g. Bregman 1989). Additional substructure


at 8.2 μm is fairly common and few sources show substructure near 7.2–7.5 μm(e.g. Moutou et al. 1999; Peeters et al. 2002). The PAH bands are often perchedon top of broad emission plateaus at roughly 3.2–3.6, 6–9, 11–14 μm (note thatvarious methods are applied to distinguish between PAH bands and plateaus; seeSect. 3.1 for details). Several H ii regions and planetary nebulae (PNe) also havebroad emission between 15–20 μm with sometimes on top the 16.4 and 17.4 μmbands. This is quite different from the collection of weaker bands as seen in e.g.RNe and galaxies (e.g. Van Kerckhoven et al. 2000; Werner et al. 2004; Peeterset al. 2004b; Sellgren et al. 2007; Smith et al. 2007; Boersma et al. 2010).

These emission features are now found in almost all astronomical environmentsincluding the (diffuse) interstellar medium (ISM), the edges of molecular clouds,reflection nebulae, some T Tauri stars, several Herbig AeBe stars, H ii regions,C-rich post-AGB stars, C-rich PNe, some C-rich WR stars, supernova remnantsand, novae. In addition, they have been found throughout the Universe out toredshifts ∼3 in various type of galaxies: normal galaxies, starburst galaxies, ultra-luminous galaxies (ULIRGs) and AGNs. The PAH emission in galaxies is not justconfined to the plane or disk of the galaxy but can also be found in their halo (e.g.Engelbracht et al. 2006; Irwin et al. 2007).

3 Spectral variations

To first order, the PAH bands are remarkably similar notwithstanding the largevariety of environments in which they are observed. However, it is clear that thePAH bands show variations in peak positions, shapes and (relative) intensities fromsource to source and also spatially within extended sources. In the analysis of thePAH bands, various assumptions are made with respect to the characteristics ofthe different components contributing to the IR emission. Generally, a featurelessdust continuum is assumed. However, different methods are applied with regard tothe PAH bands themselves. A first method assumes the broad plateaus underlyingthe individual PAH features to be part of these features. In this case, both plateausand individual features are fitted together with Lorentzians or Drude profiles (e.g.Boulanger et al. 1998; Smith et al. 2007). Alternatively, the plateaus and thefeatures are considered to be independent and local spline continua are used toseparate the components (Fig. 1, e.g. Hony et al. 2001; Peeters et al. 2002).Sometimes, the features are then fitted with Gaussian profiles. Neither approachis technically correct (see Tielens 2008 for a detailed discussion). Although thesemethods result in different PAH band strengths and profiles, the global trends inspectral variations are independent of the applied method. Some key results arediscussed below.

3.1 Intensity variations

Strong variations are observed in the relative strength of the CC and CH in-plane-bending modes in the 6 to 9 μm range relative to the CH modes at 3.3 and11.2 μm (Fig. 2, e.g. Hony et al. 2001; Galliano et al. 2008). Specifically, the


-0.4 -0.2 0.0 0.2log10(I11.2/I6.2)

-1.0

-0.8

-0.6

-0.4

log 1

0(I 3

.3/I 6

.2)

Fig. 2. Left: the correlation of the 3.3 and 11.2 μm band strengths normalized to the

6.2 μm band strength. Hexagons are H ii regions, stars intermediate mass star-forming

regions, squares RNe and triangles are PNe (Figure taken from Hony et al. 2001). Right:

the variation in the CC/CH band ratios within the starburst galaxy M 82. The gray filled

area represent the correlation obtained for the integrated spectra of a large sample of

galactic and extragalactic sources. The white (gold) symbol is the value of the global

measurement over the entire galaxy (Figure taken from Galliano et al. 2008).

strength of the 3.3 μm feature correlates with that of the 11.2 μm feature but theyboth vary considerably with respect to the strengths of the 6.2, 7.7 and 8.6 μmfeatures. Likewise, the latter three PAH bands correlate well with each other. Notehowever that the CH out-of-plane (CHoop) bending modes (10–15 μm region) donot behave consistently: the 12.7 μm PAH band correlates well with the 6 to 9 μmmodes and not with the 11.2 μm PAH band. These relative intensity variationsare widespread: they are seen from source to source and spatially within (extra-)galactic sources (Fig. 2). Combined, these studies reveal the interrelationshipbetween the various PAH bands.

PAH ratios in Galactic and Magellanic Cloud H ii regions seem to depend onmetallicity (Vermeij et al. 2002; LeBouteiller et al. in prep.). A similar depen-dence is however not observed towards Galactic and Magellanic Cloud planetarynebulae (Bernard-Salas et al. 2009). In addition, several galaxies harboring anAGN exhibit very weak PAH emission in the 6 to 9 μm region relative to the11.2 PAH band (Smith et al. 2007; Bregman et al. 2008; Kaneda et al. 2005,elsewhere in this volume). Moreover, for these galaxies, the 7.7/11.2 PAH ratiocorrelates with the hardness of the radiation field. This is in contrast to galaxieswith H ii region or starburst-like characteristics where the 7.7/11.2 PAH ratio islargely insensitive to the hardness of the radiation field. Even though, the latterinclude low metallicity systems which typically have a harder radiation field.

The total PAH intensity is highly variable and is generally studied relative tothe strength of the dust emission. The latter is commonly determined by either


the dust continuum emission at the wavelength of the considered PAH band, thedust emission at wavelengths around 15–25 μm denoted as Very Small Grains(VSGs), the far-IR (FIR) dust emission or the total IR (TIR) dust emission.The ratio PAH/VSG varies spatially across extended Galactic H ii regions suchas e.g. the Orion Bar and M 17. In particular, the dust emission dominatesinside the H ii region while in the PDR the PAHs exhibit their peak intensityand the dust continuum at 15–25 μm decreases in strength (due to its lowertemperature). Similarly, the PAH/VSG ratio in galaxies depends on the hardnessof radiation field and the metallicity (e.g. Madden et al. 2006; Brandl et al. 2006;Engelbracht et al. 2006; Gordon et al. 2008). These two parameters are relatedand so it is hard to distinguish between an origin in a less efficient PAH formationprocess (metallicity effect) or in an increased PAH processing (modification and/ordestruction of PAHs by the hard radiation field). For a detailed discussion, we referto Calzetti, Hunt, Galliano and Sandstrom, elsewhere in this volume.

3.2 Profile variations

From the early days in PAH research, it was realized that i) the 7.7 μm complexis comprised of two components at 7.6 and 7.8 μm which have variable relativestrength (Bregman 1989; Cohen et al. 1989) and ii) the 3.3 μm PAH band profileshowed small variations (e.g. Tokunaga et al. 1991). However, it is the largeamount of PAH spectra provided by ISO and subsequently Spitzer that allowed asystematic study of the PAH band profiles.

These PAH band profiles show pronounced variability, in particular for the CCmodes (6.2 and 7.7 μm bands). The variations of the band profiles for the mainPAH bands are classified in three classes A, B and C (see Fig. 3; Peeters et al.2002; van Diedenhoven et al. 2004). This classification is primarily based uponthe peak position of the bands with class C being redshifted from class B which inturn is redshifted from class A. Class C band profiles are significantly different fromclass A and B. In particular, instead of a 7.7 μm complex with either a dominant7.6 μm component (class A) or a dominant component peaking between 7.8 and8 μm (class B), class C objects show a very broad band peaking at ∼8.2 μm witha weak to absent 8.6 μm PAH band. It should also be emphasized that whileclass A and C show little variation in their profiles, large differences are presentwithin class B. Some objects have band profiles that encompasses 2 classes (AB:Van Kerckhoven 2002; Boersma et al. 2008; BC: Sloan et al. 2007) and hencethese three classes are not suggestive of three independent groups. Rather, theobserved variations in the UIR spectra seem to span a continuous distributiongoing from class A on one extreme via class B to class C on the other extreme.The observed contrast in the magnitude of the spectral variations for the CH modesversus the CC modes is striking: the peak wavelengths of the features attributedto CC modes (6.2 and 7.7 μmPAH bands) vary by ∼25 to 50 cm−1, while thevariations are smaller for the CH modes (3.3, 8.6 and 11.2 μm PAH bands; ∼ 4to 11 cm−1). In addition, although the classification was applied to each bandindividually, typically all bands in the 6–9 μm region of a single object belong


Fig. 3. The source to source variations in position and profile of the main PAH bands.

In particular large variations are evident in the 6 to 9 μm. Class A peaks at the shortest,

“nominal” wavelengths and class B at longer wavelengths. The CC modes of class C peak

at even longer wavelengths (Peeters et al. 2002; Figure taken from van Diedenhoven et al.

2004).

to the same class while the CH modes at 3.3 and 11.2 μm appear somewhat lessconnected to each other and/or to the PAH class in the 6–9 μm region.

As already noted by Bregman (1989) and Cohen et al. (1989) for the 7.7 μmcomplex, the specific profile of these PAH bands depends directly on the type ofobject (Peeters et al. 2002; van Diedenhoven et al. 2004). Class A representsH ii regions, non-isolated Herbig AeBe stars, some PNe, few post-AGB stars, RNe,the (diffuse) ISM and entire galaxies; class B contains most planetary nebulae, iso-lated Herbig AeBe stars and some post-AGB stars and class C is mainly comprisedof post-AGB stars. Few HAeBe and TTauri stars are reported to have class C pro-files but their PAH bands are weak and perched on top of both the dust continuumand silicate emission band. Note that post-AGB stars are found in all three classes.From a different perspective, class A profiles are associated with interstellar ma-terial (ISM) while class B and C profiles are exhibited by circumstellar material


Fig. 4. Left: the central wavelengths of the 7.7–8.2 and 11.3 μm PAH features plotted

versus the effective temperature of the host star (Sloan et al. 2007, Figure taken from

Keller et al. 2008). Right: empirical calibration of the 6.2/11.2 band ratio as a function

of the ionization parameter, G0T1/2/ne (Figure taken from Galliano et al. 2008).

(CSM). These spectral variations are also seen in the Magellanic Clouds (Bernard-Salas et al. 2009). Furthermore, they have also been observed within spatiallyextended objects such as RNe and evolved stars (Bregman & Temi 2005; Songet al. 2007).

Sloan et al. (2007) found a remarkable anticorrelation of the band peak posi-tion with effective temperature of the exciting star for a sample of post-AGB starsand isolated HAeBe stars, all of class B or C (Fig. 4). This relation does not seemto hold in general because: i) PNe (belonging to class A and B) can have higheffective temperatures and class B profiles (e.g. NGC 7027), ii) class A objects donot follow the correlation. For example, RNe have class A profiles and a centralstar with a low effective temperature and iii) the PAH bands of HAeBe stars withthe same effective temperature can belong to either class A or B (Van Kerckhoven2002; Boersma et al. 2008). Sloan et al. (2007) attributed this relation to differ-ent degrees of UV processing and argue that PAHs in reflection nebulae may havebeen exposed to a stronger radiation field explaining their class A PAH bands. Onthe other hand, the exceptions may also suggest that in addition to or in contrastwith a dependence on effective temperature, other parameters such as e.g. history,environment (ISM vs. CSM), spatial structure (disk, collapsing cloud; Boersmaet al. 2008) may play a role.


4 Spectral decomposition

A different and potentially powerful approach to the analysis of PAH observationsis the mathematical decomposition of PAH spectra as a linear combination of abasis set of components (Boissel et al. 2001; Rapacioli et al. 2005, elsewhere inthis volume; Joblin et al. 2008; Berne et al. 2007, 2009, elsewhere in this volume).This method extracts spatially different components from the observations. Orig-inally applied to RNe and edges of clouds, these studies found a basis set of threecomponents – dubbed VSG, PAH0 and PAH+ – with the VSG located deeper intothe PDR while the PAH0 and PAH+ are found closer to the edge of the PDR.The VSG component is characterized by broad emission bands at wavelengthsslightly redshifted to those typical for PAH emission on top of a continuum andis attributed to PAH clusters. Hence, this VSG component is not the same as theVSGs discussed earlier. The components PAH0 and PAH+ do not show continuumemission and are attributed to respectively neutral and ionized PAHs. This neutralPAH component does however exhibit stronger emission in the 6 to 9 μm regionthan expected from experimental and theoretical PAH spectra. This may indicatethe inclusion of some ionized PAH emission in this component. An alternative ex-planation may be found in the PAH size distribution. In order to fit PAH spectraof class B and C, this basis set has been expanded to include templates exhibitinga very broad band at 8.2, 8.3 and 12.3 μm (thus mimicking class C) and a thirdPAH template, PAHx, with peak positions consistent with class B. The latter isattributed to large ionized PAHs and is consistent with theoretical PAH spectra(see Sect. 5). These studies suggest the destruction of the VSGs or PAH clustersby UV photons resulting in the formation of PAHs. These PAH clusters may bereformed in the denser and more shielded environments of molecular clouds. Inregions of high UV radiation, the PAHx component gains in importance due tothe destruction of small PAHs and the higher excitation conditions.

5 Astronomical implications

The interpretation of the astronomical PAH spectra strongly relies on theoreticaland experimental spectroscopy of PAHs and related species. These studies revealthat intrinsic PAH spectra are determined by parameters such as the PAHs charge,size, the precise molecular (edge) structure and, temperature. Hence, a comparisonof astronomical spectra with laboratory and theoretical studies allows to constrainthe characteristics of the carriers and to determine the origin of the observedspectral variations. Combined with the observed behavior of PAHs in space, thisgives insight into the properties of the astronomical PAH family which can be tiedto specific environmental parameters. Here we name a few.

5.1 The astronomical & the intrinsic PAH spectra: A happy marriage

One of the strongest criticisms against the PAH hypothesis has been the lack ofa good match between a combination of experimental and/or theoretical PAH


spectra and the astronomical PAH spectra. This is now neutralized by the goodfits obtained to the three classes A, B and C with the full theoretically calculatedNASA Ames PAH spectral database5 (Figs. 1b and 2 of Cami elsewhere in thisvolume). These fits can be broken down according to size, charge and compositionunveiling the underlying origin of the different classes. For example, the fit to classC is dominated by emission of small PAHs (number of C-atoms, NC , ≤ 30).

5.2 The PAH band profiles and the composition of the PAH family

The variations in band profiles reflect primarily a (chemical) modification of thePAH family in different environments. Currently, the precise process is not agreedupon and is further constrained by the following observations. First, the modifi-cation of the PAH family needs to be consistent with the observed PAH classes atvarious stages in the PAH life cycle. Second, the position of the class A 6.2 μmPAH is not reproduced by the strongest pure CC stretching mode of pure PAHs(Peeters et al. 2002; Hudgins et al. 2005).

To resolve the second issue, various PAH-related species have been proposed.i) Hetero-atom substituted PAHs (e.g. Peeters et al. 2002; Hudgins et al. 2005;Bauschlicher et al. 2009). Substitution of a C atom by a N atom in a PAH,i.e. PANHs, systematically shifts the position of the strongest pure CC stretchingmode towards shorter wavelengths while having no systematic effects on the po-sition of other vibrational modes. This has also been noted in the decompositionof the above mentioned fits (Cami, this volume). ii) PAH-metal complexes (e.g.Hudgins et al. 2005; Bauschlicher & Ricca 2009; Simon & Joblin 2010; Joallandet al. 2009). These are species in which a metal atom is located either belowor above the carbon skeleton. See Simon et al. (elsewhere in this volume) forspecific structures. iii) PAH clusters (e.g. Rapacioli et al. 2005; Simon & Joblin2009). For example PAH dimers, see Rapacioli et al. (this volume) for specificstructures. iv) Carbon isotope effects (Wada et al. 2003). The presence of 13C inPAHs redshifts the position of the CC stretching modes and thus it will move theband further away from the observed astronomical class A 6.2 μm PAH position.Hence, it cannot resolve the issue.

Additional mechanisms are invoked to interpret the different classes. Onemechanism involves variations in the size distribution of the PAH family(Bauschlicher et al. 2008; 2009, Cami elsewhere in this volume). It was noticedthat small PAHs (NC ≤ 48) do emit at 7.6 μm but do not reproduce the 7.8 μmcomponent (e.g. Peeters et al. 2002). In contrast, large PAHs (54 ≤ NC ≤ 130)emit at 7.8 μm and not at 7.6 μm (Bauschlicher et al. 2008, 2009). Hence, inorder to reproduce the 7.7 μm complex, both small and large PAHs are requiredand hence the variable strength of the two components (class A vs. class B) maythen reflect a change of the PAH size distribution in different environments. Sim-ilarly, the broad class C component can be produced by small PAHs (Cami, thisvolume). The PAH classes may then reflect different size distributions of the PAH

5http://www.astrochem.org/pahdb/; Boersma et al. this volume.


family. Alternatively, the classes represent a varying importance of aliphatics vs.aromatics (e.g. Sloan et al. 2007; Boersma et al. 2008; Keller et al. 2008; Pinoet al. 2008; Acke et al. 2010). Aliphatic carbonaceous species are known toproduce broader emission bands compared to PAHs and hence are put forwardas the carrier of the class C 7.7 μm complex. An evolution to class B and classA is then obtained by increased UV processing (e.g. due to an increase in effec-tive temperature) of these aliphatic carbonaceous species destroying the aliphaticbonds and hence increasing the aromaticity of species. UV processing can alsochange the PAH size distribution and hence both mechanisms together may lay atthe origin of the profile changes. While the second mechanism (towards increasedaromaticity) can explain the evolutionary scenario from post-AGB stars to theISM, it is hard to imagine a reversed process from aromatic material to more frag-ile aliphatic material when going from the ISM to proto-planetary environments.Hence, Boersma et al. (2008) proposed an active chemical equilibrium betweenaromatic and aliphatic species in all environments through hydrogenation, carbonreactions building (aliphatic) hydrocarbons and UV processing.

5.3 The CC/CH ratio and the charge balance of PAHs

Laboratory and theoretical studies on PAHs have shown the remarkable effect ofionisation on their IR spectra (e.g. Hudgins & Allamandola 2004, Pauzat elsewherein this volume). While peak positions are only modestly affected, the influenceon intensity is striking: the bands in the 5–10 μm region grow from the small-est features to become the dominant bands upon ionization. In addition, in the10–15 μm region, the 11.0 μm PAH band can be attributed to ionized PAHs whilethe 11.2 μm band is due to neutral PAHs. Besides charge state, there are severalother parameters that may influence the CC/CH ratio (i.e. the ratio of the 6.2 or7.7 μm PAH band to the 3.3 or 11.2 μm PAH band) including dehydrogenation, amodification of the temperature distribution (due to a change in the size distribu-tion or the radiation field) and extinction. However, several studies indicate thatthe observed variation in the CC/CH ratio is dominated by a variation in the de-gree of ionization of the PAHs in these different environments (Joblin et al. 1996;Hony et al. 2001; Galliano et al. 2008 and Galliano elsewhere in this volume).

5.4 Molecular edge structure of PAHs

The peak wavelength of the CHoop bending modes depends strongly on the numberof adjacent peripheral C-atoms bonded to an H-atom (Bellamy 1958; Hony et al.2001). While the exact peak position depends slightly on the charge state, the11.0 and 11.2 PAH bands originate in solo CH groups (i.e. no H-atom is attachedto adjacent C-atoms). Both duo and trio CH groups (respectively two and threeadjacent C-atoms each with an H-atom attached) contribute to the 12.7 μm band,trio CH groups produce the 13.5 μm band and quatro CH groups the 14.2 μmband. The observed relative intensities of the CHoop bands can then be combinedwith the intrinsic strengths of these modes to determine possible PAH structures


(Hony et al. 2001; Bauschlicher et al. 2008, 2009). After all, solo-CH groups dec-orate the long, smooth straight edges while the other CH groups are characteristicsfor corners and bays in the (irregular) edge structure. PAHs associated with PNeare characterized by very compact molecular structures with long smooth PAHedges while interstellar PAHs have more corners, either because they are on aver-age smaller or they are more irregular larger species. Similarly, the CH stretchingmode of H atoms located in bay regions emit around ∼3.22 μm, a region wherelittle emission is found in astronomical PAH spectra. Therefore, the presence ofbay regions in the edge structure of (small) PAHs is very low (Bauschlicher et al.2009).

6 PAHs as a diagnostic tool

The CC/CH ratio is determined by the charge balance of the PAHs. The PAHcharge is set by the ratio of the ionization rate to the recombination rate, that isproportional to G0T1/2/ne where G0 is the UV radiation field, T the gas temper-ature and ne the electron density (Tielens 2008). Thus, by coupling the variationsin the PAH bands with known variations in the physical conditions (as derived byfor example PDR models), an empirical calibration can be established that relatesthe PAH bands to the local physical conditions. Such a calibration will allow thedetermination of the physical conditions based upon the omnipresent PAH bands.This can then serve as a diagnostic tool for regions where, for example, the mainPDR coolants are not easily observable, such as galaxies at large distances. Thisapproach has been tested for a sample of three well studied objects and has beenproven very promising (Fig. 4, Galliano et al. 2008; Galliano in this volume). Sim-ilarly, combining the observed PAH ionization ratio and H2 line ratios in dense,highly irradiated PDRs allows to derive the physical conditions (Berne et al. 2009,elsewhere in this volume).

Since PAHs are excited by UV radiation, PAH features are also particularlybright in massive star-forming regions. Combined with their omnipresence, thismake PAHs a powerful tracer of star formation throughout the Universe (Calzettiand Hunt et al. in this volume and references therein). They are widely used toderive star formation rates of galaxies. And in combination with emission lines,they serve as diagnostics for the ultimate physical processes powering galacticnuclei (e.g. Genzel et al. 1998; Lutz et al. 1998; Peeters et al. 2004c; Sajinaet al. 2007; Smith et al. 2007). Finally, the presence of PAHs is used to distinguishbetween shocked gas and PDRs (van den Ancker et al. 2000) and to determineredshifts in distant galaxies (e.g. Yan et al. 2007).

7 Future

This paper reviews the large progress made over the past 25 years in understand-ing the unidentified infrared bands. Nevertheless, several key questions remainregarding astronomical PAHs including: How do the PAH characteristics (e.g.


size, charge state) interact with and reflect the physical conditions of their envi-ronment (e.g. density, radiation field, temperature, metallicity)?; Do we observelong wavelength counterparts to the well-known mid-IR PAH bands? What aretheir characteristics?; How can we use the UIR bands as a probe of the phys-ical conditions in regions near and far?; Which specific molecules make up theastronomical PAH family?

The observational future for PAH research is very bright. Currently, we are for-tunate to obtain the necessary observations with the Herschel Space Observatoryto explore the far-IR modes of the PAHs and to couple the PAH band character-istics to the physical conditions of the environment (cf. Joblin et al., elsewhere inthis volume). In addition, in the near future, we will use SOFIA and the JamesWebb Space Telescope (JWST). But the wealth of PAH emission bands and thelarge variability of the PAH spectrum already prompt more questions than can beanswered with the current laboratory and theoretical data. If we intend to makesignificant progress in our understanding of the PAH bands, the observationaleffort needs to be balanced with dedicated laboratory and theoretical studies. In-deed, we can only fully exploit the treasure of information hidden in the PAHemission bands by a joint effort of the observational, experimental and theoreticaltools.

References

Acke, B., Bouwman, J., Juhasz, A., et al., 2010, ApJ, 718, 558

Barker, J.R., Allamandola, L.J., & Tielens, A.G.G.M., 1987, ApJ, 315, L61

Bauschlicher, C.W.Jr, Peeters, E., & Allamandola, L.J., 2008, ApJ, 678, 316

Bauschlicher, C.W.Jr, Peeters, E., & Allamandola, L.J., 2009, ApJ, 697, 311

Bauschlicher, C.W.Jr, & Ricca, A., 2009, ApJ, 698, 275

Bellamy, L., 1958, The infra-red spectra of complex molecules, 2nd ed. (New York: JohnWiley & Sons, Inc.)

Bernard-Salas, J., Peeters, E., Sloan, G.C., et al., 2009, ApJ, 699, 1541

Berne, O., Joblin, C., Deville, Y., et al., 2007, A&A, 469, 575

Berne, O., Fuente, A., Goicoechea, J.R., et al., 2009, A&A, 706, L160

Boersma, C., Bouwman, J., Lahuis, F., et al., 2008, A&A, 484, 241

Boersma, C., Bauschlicher, C.W.Jr, Allamandola, L.J., et al., 2010, A&A, 511, 32

Boissel, P., Joblin, C., & Pernot, P., 2001, A&A, 373, L5

Boulanger, F., Boisssel, P., Cesarsky, D., & Ryter, C., 1998, A&A, 339, 194

Brandl, B.R., Bernard-Salas, J., Spoon, H.W.W., et al., 2006, ApJ, 653, 1129

Bregman, J. 1989, in IAU Symp. 135: Interstellar Dust, ed. Allamandola L.J., & Tielens,A.G.G.M., 109

Bregman, J., & Temi, P., 2005, A&A, 621, 831

Bregman, J.D., Bregman, J.N., & Temi, P., 2008, ASPC, 381, 34

Cami, J., Bernard-Salas, J., Peeters, E., & Malek, S.E., 2010, Science, 329, 1180

Cohen, M., Tielens, A.G.G.M., & Bregman, J., et al., 1989, ApJ, 341, 246

Engelbracht, C.W., Kundurthy, K.D., Gordon, K.D., et al., 2006, ApJ, 642, L127


Galliano, F., Madden, S., Tielens, A.G.G.M., et al., 2008, ApJ, 679, 310

Geballe, T.R., Lacy, J.H., Persson, S.E., McGregor, P.J., & Soifer, B.T., 1985, ApJ, 292,500

Genzel, R., Lutz, D., Sturm, E., et al., 1998, ApJ, 498, 579

Gillett, F.C., Forrest, W.J., & Merril, K.M., 1973, ApJ, 183, 87

Gordon, K.D., Engelbracht, C.W., Rieke, G.H., et al., 2008, ApJ, 682, 336

Hony, S., Van Kerckhoven, C., Peeters, E., et al., 2001, A&A, 370, 1030

Hudgins, D.M., Bauschlicher, C.W. Jr., & Allamandola, L.J., 2005, ApJ, 632, 316

Hudgins, D.M., & Allamandola, L.J., 2004, Astrophys. Dust, ASP Conf. Ser., 309, 665

Irwin, J.A., Kennedy, H., Parkin, T., & Madden, S., 2007, A&A, 474, 461

Joalland, B., Simon, A., Marsden, C.J., & Joblin, C., 2009, A&A, 494, 969

Joblin, C., Szczerba, R., Berne, O., & Szyszka, C., 2008, A&A, 490, 189

Joblin, C., Tielens, A.G.G.M., Geballe, T.R., & Wooden, D.H., 1996, ApJ, 460, L119

Kaneda, H., Onaka, T., & Sakon, I., 2005, ApJ, 632, L83

Keller, L.D., Sloan, G.C., Forrest, W.J., et al., 2008, ApJ, 684, 411

Kennicutt, R.C.Jr., Armus, L., Bendo, G., et al., 2003, PASP, 115, 928

Lutz, D., Spoon, H.W.W., Rigopoulou, D., Moorwood, A.F.M., & Genzel, R., 1998, ApJ,505, L103

Madden, S., Galliano, F., Jones, A.P., & Sauvage, M., 2006, A&A, 446, 877

Moutou, C., Sellgren, K., Verstraete, L., & Leger, A., 1999, A&A, 347, 949

Pech, C., Joblin, C., & Boissel, P., 2002, A&A, 388, 639

Peeters, E., Hony, S., van Kerckhoven, C., et al., 2002, A&A, 390, 1089

Peeters, E., Allamandola, L.J., Bauschlicher, C.W.Jr, et al., 2004a, ApJ, 604, 252

Peeters, E., Mattioda, A.L., Hudgins, D.M., & Allamandola L.J., 2004b, ApJ, 617, L65

Peeters, E., Spoon, H.W.W., & Tielens, A.G.G.M., 2004c, ApJ, 613, 986

Pino, T., Dartois, E., Cao, A.-T., et al., 2008, A&A 490, 665

Rapacioli, M., Joblin, C., & Boissel, P., 2005, A&A, 429, 193

Roche, P.F., Lucas, P.W., & Geballe, T.R., 1996, MNRAS, 281, L25

Sajina, A., Yan, L., & Armus, L., et al., 2007, ApJ, 664, 713

Sellgren, K., Uchida, K.I., & Werner, M.W., 2007, ApJ, 659, 1338

Sellgren, K., Werner, M.W., & Ingals, J., et al., 2010, ApJ, 722, L54

Simon, A., & Joblin, C., 2009, J. Phys. Chem. A, 113, 4878

Simon, A., & Joblin, C., 2010, ApJ, 712, 69

Sloan, G.C., Jura, M., Duley, W.W., et al., 2007, ApJ, 664, 1144

Smith, J.D., Draine, B.T., Dale, D.A., et al., 2007, ApJ, 656, 770

Song, I.-O., McCombie, J., Kerr, T.H., & Sarre, P.J., 2007, MNRAS, 380, 979

Tielens, A.G.G.M., 2008, ARA&A, 46, 289

Tokunaga, A.T., Sellgren, K., Smith, R.G., et al., 1991, ApJ, 380, 452

van den Ancker, M.E., Tielens, A.G.G.M., & Wesselius, P.R., 2000, A&A, 358, 1035

van Diedenhoven, B., Peeters, E., Van Kerckhoven, C., et al., 2004, ApJ, 611, 928

Van Kerckhoven, C., 2002, Ph.D. Thesis, Catholic University of Leuven

Van Kerckhoven, C., Hony, S., Peeters, E., et al., 2000, A&A, 357, 1013


Vermeij, R., Peeters, E., Tielens, A.G.G.M., & van der Hulst, J.M., 2002, A&A, 382,1042

Verstraete, L., Pech, C., Moutou, C., et al., 2001, A&A, 372, 981

Wada, S., Onaka, T., Yamamura, I., Murata, Y., & Tokunaga, A.T., 2003, A&A, 407,551

Werner, M.W., Uchida, K.I., Sellgren, K., et al., 2004, ApJS, 154, 309

Yan, L., Sajina, A., Fadda, D., et al., 2007, ApJ, 658, 778


ASTRONOMICAL MODELS OF PAHS AND DUST

B.T. Draine1

Abstract. Physical dust models which include a PAH component arequite successful in reproducing the measured extinction vs. wavelengthin the Milky Way, the infrared (IR) emission observed from Milky Wayregions and other galaxies, and a number of other observational con-straints. Many of the adopted PAH properties are necessarily highlyidealized.

The observed variations in the 7.7 μm/11.3 μm band ratio can bereproduced by varying the PAH ionization balance. Changing the spec-trum of the starlight heating the PAHs affects the overall strength ofthe PAH emission (relative to total IR), but has relatively little effecton the 7.7 μm/11.3 μm band ratio.

1 Introduction

PAHs are an important component of the dust population. They make a sub-stantial contribution to the overall extinction by dust, accounting for ∼20% of thetotal IR power from a star-forming galaxy like the Milky Way. See Tielens (2008)for a general review of PAHs.

Physical models for dust, including a PAH population, have been developed(Desert et al. 1990; Weingartner & Draine 2001; Li & Draine 2001a; Zubko et al.2004; Draine & Li 2007; Draine & Fraisse 2009; Compiegne et al. 2010). Forapplication to the Milky Way, these models are subject to an array of observationalconstraints:

• Starlight extinction: The observed extinction of starlight strongly constrainsthe dust size distribution and composition. Extinction curves have beenmeasured using stars in the Milky Way (MW), Large Magellanic Cloud(LMC), and Small Magellanic Cloud (SMC), and in a limited number ofother galaxies using background quasistellar objects (QSOs) and gamma rayburst (GRB) afterglows as the light source. The extinction is best-studied

1 Dept. of Astrophysical Sciences, Princeton University



in the MW, where spectral features at 10 μm and 18 μm show that amor-phous silicates are a major constituent, and a strong extinction “bump” near2175 A is usually interpreted as due to sp2-bonded carbon, as in graphite orPAHs.

• Starlight polarization: The extinction is polarization-dependent, requiringthat some of the grains be nonspherical and aligned with respect to theGalactic magnetic field.

• Scattering of starlight: A substantial fraction of the observed extinction inthe optical is due to scattering, requiring that some of the grains must belarge enough to efficiently scatter optical light.

• Abundance constraints: The grain model should incorporate elements onlyas allowed by the total abundance of the element minus the fraction observedto be in the gas phase.

• Infrared emission: The dust model should reproduce the IR emission spectraobserved from regions with different intensities of starlight heating the dust.

• X-Ray scattering: The dust composition and size distribution must be suchas to reproduce the observed strength and angular distribution of X-rayscattering by interstellar dust.

• Microwave emission: Dust-correlated microwave emission is attributed torotational emission from very small dust grains (Draine & Lazarian 1998a,b).This constrains the abundances of the smallest grains.

• Dust in meteorites: Presolar grains with isotopic anomalies are found inmeteorites. These were part of the interstellar grain population 4.6 Gyr ago,when the solar system formed.

Models for MW dust are therefore strongly constrained (see Draine 2003, for fur-ther discussion).

2 Physical models for dust with PAHs

The Infrared Space Observatory (ISO) discovered, and Spitzer Space Telescopeconfirmed, that strong PAH emission features are routinely present in the IRspectra of star-forming galaxies. The relative strengths of the PAH features dovary somewhat from galaxy to galaxy, but in fact show surprising “universality”.

Measured extinction curves in the MW, LMC, and SMC have many similarities:(1) A general trend to rise strongly from the near-IR (λ ≈ 1 μm) to the FUV(λ ≈ 0.1 μm). (2) A tendency for AV /NH to vary in proportion to metallicity –the MW, LMC, and SMC all seem to have similar fractions of “refractory elements”in dust. Models to reproduce the observed Aλ/NH require a substantial fractionof abundant refractory elements (C, Mg, Si, Fe).

B.T. Draine: Astronomical Models of PAHs and Dust 31

However, significant regional variations in the shape of the extinction curve areobserved. The function Aλ/AV varies from sightline-to-sightline within the MW,and within the LMC and SMC. In particular, the relative strength of the 2175 Afeature decreases as one goes from MW to LMC to SMC. A physical grain modelshould be able to accomodate these variations.

Contemporary dust models for dust in the MW (and other galaxies) (e.g.,Weingartner & Draine 2001; Li & Draine 2001b; Zubko et al. 2004; Draine &Li 2007; Draine & Fraisse 2009) have amorphous silicate and carbonaceousmaterial as the principal dust materials. The specific mix of carbonaceous mate-rial (PAH, graphite, various forms of amorphous carbon...) varies from model tomodel.

A dust model consists of specifying the grain composition and size distribution.The MW, LMC, and SMC extinction curves can be reproduced by models con-sisting of amorphous silicate + graphite + PAHs, with only changes in the sizedistributions (and relative abundances) of the 3 components from sightline tosightline (Weingartner & Draine 2001). In galaxies lacking measured extinctioncurves, it is reasonable to use the size distribution obtained for the MW unless weare forced to change some property, such as the relative abundance of PAHs.

For each grain composition and size, we require scattering and absorption crosssections as a function of wavelength. For grains with radii a > 100 A, we calculatethe scattering and absorption cross sections by solving Maxwell’s equations for theappropriate grain size and dielectric function, usually assuming either a sphericalor spheroidal shape (Draine & Fraisse 2009). To calculate the response to single-photon heating, we also require the heat capacity of each grain.

The PAHs are small enough that scattering is expected to be unimportant,and are also small enough that calculating absorption cross sections using “bulk”optical constants – even if these were available – may be a poor approximation.Instead, we rely on experimental and theoretical studies of PAH absorption crosssections per C atom. The PAH opacity adopted by Draine & Li (2007, hereafterDL07) is shown in Figure 1.

Our knowledge of absorption cross sections for PAHs (neutrals and ions) is lim-ited, especially as regards large species (cf. http://astrochemistry.ca.astro.it/database/pahs.html). At this time most dust models make the simplify-ing assumption that the band profiles for PAHs are “universal”. Band positions,widths, and strengths are adopted that appear to be consistent with (1) astro-nomical observations of band profiles, and (2) laboratory and theoretical studies(to the extent available). In reality, cross sections vary significantly from PAH toPAH, and from region to region (see Peeters 2011, in this volume) but we are notyet in a position to include this in models. Ionization matters: the adopted bandstrengths depend on whether the PAH is neutral or ionized. A first study on therole of charge state of PAHs in ultraviolet extinction has been recently performed(Cecchi-Pestellini et al. 2008; see Mulas et al. 2011 in this volume).


Fig. 1. Radiative cross sections adopted for PAHs (neutral and cations) in the UV-optical

(left) and IR (right). After Draine & Li (2007).

3 Modeling the IR emission

In the ISM, grains and PAHs are heated primarily by absorption of starlight pho-tons, and cooled by emission of IR photons. Radiative deexcitation is a quantumprocess, but Aν(E)dν, the probability per unit time of spontaneous emission of aphoton in [ν, ν + dν] by a PAH with internal energy E, can be approximated by

Aν(E) ≈ 4π

hνCabs(ν)Bν(T (E))

where Bν(T ) = blackbody function, and T (E) is such that an ensemble of suchPAHs at temperature T would have average vibrational energy equal to E. Thethermal approximation is valid except at the lowest vibrational energies(Allamandola et al. 1989; Leger et al. 1989; Draine & Li 2001).

For small PAHs, the time between photon absorptions is in general long com-pared to the time required for the PAH to deexcite by IR emission, and thereforethe PAH temperature T undergoes large excursions. For larger particles, and in-tense radiation fields, the grain is unable to fully cool between photon absorptions.The PAH temperature distribution function is found by solving the equations ofstatistical equilibrium, with upward transitions due to photon absorptions, anddownward transitions due to spontaneous emission. Figure 2 shows examples oftemperature distribution functions for PAH ions.

The temperature distribution function depends on the PAH size, and thereforethe total emitted spectrum will be sensitive to the PAH size distribution dn/da.Figure 3 shows dn/da adopted by DL07 for diffuse clouds in the MW, as well as thatused by Zubko et al. (2004). While both have similar amounts of carbonaceous


Fig. 2. Temperature distribution functions for PAH+ particles in (a) the standard ISRF,

and (b) a radiation field 104 times more intense than the standard ISRF. Curves are

labelled by the radius of an equal-volume sphere. After Draine & Li (2007).

material in particles with < 103 C atoms, the size distributions differ in detail.The differences can be taken as an indication of the uncertainties.

PAHs will in general be rotationally excited. They are expected to have nonzeroelectric dipole moments, and, therefore, to emit electric dipole radiation in rota-tional transitions (see Verstraete 2011 in this volume). Draine & Lazarian (1998a)proposed that the PAHs could account for the dust-correlated microwave emis-sion discovered by COBE (Kogut et al. 1996). The angular momentum quantumnumber J >∼ 102, allowing a largely classical treatment of the rotational excitationand damping (Draine & Lazarian 1998b). The rotation rate depends on the PAHsize. For the size distribution used to account for the PAH emission features in theinfrared results, the predicted rotational emission appears to be in agreement withthe observed intensity of dust-correlated microwave emission (e.g., Dobler et al.2009; Ysard et al. 2010). There do not appear to be any mechanisms that caneffectively align PAH angular momenta with the galactic magnetic field (Lazarian& Draine 2000), and the rotational emission from PAHs is therefore expected to beessentially unpolarized, in agreement with upper limits on the polarization of thedust-correlated microwave emission (Battistelli et al. 2006; Mason et al. 2009).By contrast, the far-infrared and submm emission from larger grains is expected tohave polarizations as large as ∼10% (Draine & Fraisse 2009) – soon to be measuredby Planck.

Recent theoretical studies have refined the treatment of the rotational dynamicsof small particles (Ali-Haımoud et al. 2009; Hoang et al. 2010; Silsbee et al. 2010;Ysard & Verstraete 2010). The evidence to date supports the view that the PAH


Fig. 3. Volume distributions ∝ a3dn/d ln a for small carbonaceous particles. a is the

radius of an equal-volume sphere; the number of carbon atoms NC is shown at the top of

the figure. The curve labeled DL07 is the size distribution used by Draine & Li (2007);

the curves labeled ZDA04 are for graphite and PAH particles from Zubko et al. (2004).

population is responsible for the bulk of the observed dust-correlated microwaveemission.

4 PAH ionization balance

The PAH charge state changes due to collisions with electrons, collisions with ions,and photoelectric emission. Rates for these processes can be estimated, and thesteady-state charge distribution function can be solved for as a function of sizeand environment. Figure 4 shows the calculated φion, the fraction of PAHs thatare non-neutral, as a function of size. The solid curve is a weighted sum over thethree idealized environments.

5 Comparison of model and observed spectra

The DL07 dust model has been applied to try to reproduce observed IR emis-sion from galaxies. The model is found to reproduce the broad-band photometryfrom galaxies both globally (e.g., Draine et al. 2007) and within galaxies (e.g.,Munoz-Mateos et al. 2009). How well does the model compare with observedspectra?

Figure 5 shows the low-resolution spectroscopy and broadband photometrymeasured in the central few kpc of SINGS galaxies (Smith et al. 2007), together


Fig. 4. Calculated ionized fraction φion for PAHs, as a function of PAH size. Solid

curve is φion adopted by DL07 for the overall ISM. From Li & Draine (2001a).

with the spectrum of a best-fit DL07 dust model, where the total dust mass, thePAH abundance (as a fraction of the total grain mass), and the distribution ofstarlight intensities have been adjusted. The agreement is impressive, particularlysince two things that were not varied were (1) the PAH ionization fraction φion(a)(held fixed at the distribution originally estimated to apply to the Milky Way –see Fig. 4), and (2) the spectrum of the starlight heating the dust (taken to be theMathis et al. 1983, hereafter MMP83) ISRF spectrum.

6 Spectral variations

While the overall agreement between model and observations in Figure 5 is excel-lent, close scrutiny discloses differences. In NGC 5195, the observed ratio of the7.7 μm feature to the 11.3 μm feature falls below the model; for NGC 6946 theobserved 7.7/11.3 ratio is slightly higher than the model.

A systematic study of variations in PAH spectra was carried out by Gallianoet al. (2008). Figure 6a shows examples of variations in the 5–16 μm PAH spectrafrom galaxy to galaxy, and within M 82. The ratio of the power in the 7.7μmcomplex to the power in the 11.3 μm feature varies considerably. Figure 6b showsthat while I(7.7)/I(11.3) varies by up to a factor 6, I(7.7)/I(6.2) and I(8.6)/I(6.2)remain nearly constant.


Fig. 5. Spectra of the central regions of SINGS galaxies (Smith et al. 2007), reproduced

by models of starlight plus IR emission from dust. Numerous (very small) data points

with error bars are 5.3 – 38 μm IRS observations, and the rectangles are IRAC and

MIPS photometry for the region where the IRS spectrum was extracted. Solid curves are

calculated for the dust model; the triangles show the model convolved with the IRAC

and MIPS response functions. For each model the value of qPAH ≡ (the fraction of

total dust mass contributed by PAHs with NC < 103 C atoms) is given (Reyes et al., in

preparation).

Figure 6c shows that for normal galaxies the 7.7/11.3 ratio does not appear todepend on the hardness of the radiation field (using the NeIII/NeII ratio in H IIregions as a proxy). However, very low values of 7.7/11.3 can be found for someAGN with high ratios of [NeIII]15.6 μm/[NeII]12.8 μm.

Can the present dust model accomodate the observed variations in band ratios?Here we consider the effect of changing the PAH ionization, and changing thespectrum of the starlight heating the dust.


(a) from Galliano et al. (2008)

(b) from Galliano et al. (2008)

0.1 1.0 10.0L([NeIII] 15.6μm)/L([NeII] 12.8μm)

1

10

L(7.

7μm

Com

plex

)/L(

11.3

μm C

ompl

ex)

ngc4631

ngc5194

ngc5713

ngc7331

(c) from Smith et al. (2007)

Fig. 6. (a) Variations in spectra from galaxy to galaxy, and within M 82 (from Galliano

et al. 2008). (b) In a sample of 50 objects, the I(7.7 μm)/I(11.3 μm) ratio varies by a fac-

tor ∼6, while I(7.7)/I(6.2) and I(8.6)/I(6.2) remain relatively constant (from Galliano

et al. 2008). (c) Variations of I(7.7)/I(11.3) among galaxies, as a function of the

[NeIII]15.6 μm/[NeII]12.8 μm flux ratio. Squares: normal HII-type galaxies; Triangles:

AGN (from Smith et al. 2007).

PAH ionization

With the ionization-dependent opacities of Figure 1, we can alter the emissionspectrum by changing the adopted ionization function φion(a). Figure 7a showsemission spectra calculated for two extreme assumptions regarding the ionization:φion = 0 and φion = 1, as well as for φion adopted by DL07 (shown in Fig. 4).It is apparent that the 7.7/11.3 band ratio increases considerably as the PAH


Fig. 7. (a) 3–30 μm emission spectrum for fully-ionized (solid curve) and neutral PAHs

(broken curve) heated by the local ISRF (Mathis et al. 1983, MMP83). The 7.7/11.3

band ratio changes by a factor ∼6. (b) 3–30 μm emission spectrum for fixed ionization

φ(a) used by DL07, but comparing PAHs heated by the MMP83 ISRF (broken curve)

vs. PAHs heated by a 2× 104 K blackbody cut off at 13.6 eV (solid curve). The 7.7/11.3

band ratio is essentially unchanged.

Table 1. PAH band ratios calculated for the DL07 model.

φion Radiation Field (νPν)7.7 μm/(νPν)11.3 μm (νPν)7.7 μm/P (TIR) a

φion = 0 MMP83 0.487 0.323φion = 0 20 kK blackbody 0.512 0.523DL07 MMP83 1.01 0.499DL07 20 kK blackbody 1.05 0.783

φion = 1 MMP83 3.08 0.759φion = 1 20 kK blackbody 3.36 1.162

a P (TIR) =∫

Pνdν is the total IR power for the dust model.

ionized fraction is increased. Table 1 shows that for the adopted PAH0 and PAH+

absorption cross sections (see Fig. 1) the 7.7/11.3 feature ratio changes by a factor∼6.3 as the ionization is varied – a range comparable to what is actually seen (seeFig. 6b)!

Variations in starlight spectrum?

There will naturally be variations in the spectrum of the starlight heating PAHs,both regionally within a galaxy, and from one galaxy to another. Heating by aharder radiation field will lead to an increase in the fraction of the dust heatingcontributed by hard photons, and therefore will increase the short wavelengthPAH emission. However, Galliano et al. (2008) showed that the harder starlight


spectrum of a starburst does not increase the 7.7/11.3 ratio by enough to explainthe observed variations. Figure 7 compares the 3–30 μm emission spectrum forPAHs heated by the MMP83 ISRF, vs. dust heated by a 2×104 K blackbody cutoff at 13.6 eV. The 2×104 K blackbody is adjusted in intensity to give the sameoverall power per H absorbed by dust. The harder radiation field does lead toan increase in the PAH emission as a fraction of the total, but has only a smalleffect on the 7.7/11.3 band ratio (only a 4% change – see Table 1), confirming thefinding of Galliano et al. (2008).

Changes in the PAH size distribution?

What about the lowest values of 7.7/11.3 found for some AGN (see Fig. 6c)? Thesmallest PAHs may be susceptible to destruction by X-rays Voit (1992). Suppres-sion of the abundances of PAHs with < 103 C atoms (while leaving larger PAHsunchanged) would reduce the 7.7/11.3 ratio, lowering it by a factor ∼1.5 (Gallianoet al. 2008); however, as noted by Galliano et al. (2008), this change to the sizedistribution would also lower the 6.2/7.7 band ratio, which does not appear tovary in normal galaxies; hence this effect would apply only in unusual objects.

7 What is the relation between PAHs and the 2175 A feature?

The strength of the interstellar 2175 A feature requires that the carrier X havenXfX/nH ≈ 9.3 × 10−6 (Draine 1989), where fX is the oscillator strength per X .In general, PAHs have strong absorption near 2200 A due to π → π∗ electronicexcitation. For small graphite spheres, the π → π∗ excitation has an oscillatorstrength f ≈ 0.16 per C atom. If we assume a similar oscillator strength for theπ → π∗ transition in PAHs, then C/H≈ 58ppm in PAHs would be sufficient toaccount for the observed integrated strength of the 2175 A feature. The abundanceof PAHs required to explain the observed IR emission is 30–60ppm (Li & Draine2001a; Draine & Li 2007). Therefore, it seems likely that the 2175 A feature isdue mainly to PAHs (Joblin et al. 1992; Li & Draine 2001a).

The sp2-bonded carbon in PAHs also has strong absorption in σ → σ∗ transi-tions, resulting in absorption rising steeply beginning at ∼1200 A (∼10 eV) andpeaking at ∼720 A (∼17 eV). Thus, any sightline with 2175 A absorption (if dueto PAHs) should also have steeply rising extinction at λ <∼ 1200 A, although otherelements of the dust population (e.g., small silicate particles) can also contributesteeply rising far-UV extinction in addition to that provided by the PAHs.

PAH emission features are ubiquitous in IR spectra of star-forming galaxies,including starburst galaxies. However, the 2175 A feature is weak or absent inspectra of starburst galaxies: the “Calzetti extinction law” for these starburstsshows no 2175 A feature (Calzetti et al. 1994). Conroy (2010) found no evidenceof 2175 A extinction in the spectra of star-forming galaxies at z ≈ 1. Note,however, that Conroy et al. (2010) do find evidence for 2175 A extinction inGALEX observations of nearby galaxies. Given the ubiquity of PAH emission instar-forming galaxies, the weakness or absence of the 2175 A feature in the spectraof starburst galaxies, and in the z ∼ 1 sample of Conroy (2010), is surprising.


Absence of the 2175 A feature in starburst spectra may be in part the result ofcomplex geometry and radiative transfer – perhaps the regions in the galaxy thatdominate the emergent FUV may be regions where the PAHs have been destroyed.

In the SMC, 4/5 sightlines where UV extinction has been measured have nodetectable 2175 A feature (Gordon et al. 2003). If PAHs absorb near 2175 A,then we expect PAH emission to be weak in the SMC. From COBE-DIRBE ob-servations, Li & Draine (2002) argued that PAH emission was strongly suppressed(relative to total IR emission) in SMC (but this was controversial – see Bot et al.2004).

Recently, Sandstrom et al. (2010) studied PAH emission in the SMC, findingthat the PAH emission, while not zero, is indeed weak, consistent with the 2175 Aextinction feature being due to PAHs. Sandstrom et al. also found large regionalvariations in the PAH abundance within the SMC, with the PAH abundance peak-ing near molecular clouds.

8 Summary

The principal points are the following:

• Physical models for interstellar dust can reproduce the observed PAH emis-sion features together with other observed properties, including microwaveemission, submm and far-IR emission, and extinction and scattering from thenear-IR to X-ray energies. Amorphous silicate material accounts for ∼75%of the dust mass. Carbonaceous material accounts for ∼25% of the mass,with ∼5% of the total dust mass in PAHs.

• Lacking detailed knowledge of the physical properties of PAHs, current mod-els use very simplified representations of PAH absorption as a function ofwavelength. At present these properties are “tuned” to be able to reproduceastronomical observations, but the adopted opacities appear to be consistentwith what is known of PAHs in the IR through ultraviolet.

• PAH band ratios are observed to vary from region to region. The 7.7/11.3band ratio varies by up to a factor ∼6. Model calculations (see Fig. 7)show that this could result from changes in the PAH ionization balance fromnearly fully neutral (low 7.7/11.3 ratio) to nearly fully ionized (high 7.7/11.3ratio).

• The 2175 A interstellar extinction feature should be due, at least in part,to π → π∗ absorption in PAHs. Weakness or absence of the 2175 A featurewould then imply weakness or absence of PAH emission. The 2175 A featureis observed to be weak in the SMC; recent study of the PAH emission from theSMC confirms that, as predicted, the PAH abundance is very low (Sandstromet al. 2010).

• Existing models use highly simplified descriptions of the PAH absorptioncross sections from the UV to the IR, with considerable freedom for adjustment


to reproduce astronomical observations. As both laboratory and theoreti-cal understanding moves forward, modeling will become more strongly con-strained.

I am grateful to Christine Joblin and the SOC for organizing this meeting, and for the invitationto speak at it. I thank Xander Tielens for comments and suggestions that helped improve thistext. This work was supported in part by NASA through JPL grant 1329088.

References

Ali-Haımoud, Y., Hirata, C.M., & Dickinson, C., 2009, MNRAS, 395, 1055

Allamandola, L.J., Tielens, G.G.M., & Barker, J.R., 1989, ApJS, 71, 733

Battistelli, E.S., Rebolo, R., Rubino-Martın, J.A., et al., 2006, ApJ, 645, L141

Bot, C., Boulanger, F., Lagache, G., Cambresy, L., & Egret, D., 2004, A&A, 423, 567

Calzetti, D., Kinney, A.L., & Storchi-Bergmann, T., 1994, ApJ, 429, 582

Cecchi-Pestellini, C., Malloci, G., Mulas, G., Joblin, C., et al., 2008, A&A, 486, L25

Compiegne, M., Verstraete, L., Jones, A., et al., 2010 [ArXiv 1010.2769v1]

Conroy, C., 2010, MNRAS, 404, 247

Conroy, C., Schiminovich, D., & Blanton, M.R., 2010, ApJ, 718, 184

Desert, F.-X., Boulanger, F., & Puget, J.L., 1990, A&A, 237, 215

Dobler, G., Draine, B., & Finkbeiner, D.P., 2009, ApJ, 699, 1374

Draine, B.T., 1989, ed. L. Allamandola & A. Tielens, IAU Symp. 135: Interstellar Dust(Kluwer, Dordrecht), 313

Draine, B.T., 2003, ARAA, 41, 241

Draine, B.T., Dale, D.A., Bendo, G., et al., 2007, ApJ, 663, 866

Draine, B.T., & Fraisse, A.A., 2009, ApJ, 696, 1

Draine, B.T., & Lazarian, A., 1998a, ApJ, 494, L19

Draine, B.T., & Lazarian, A., 1998b, ApJ, 508, 157

Draine, B.T., Li, A., 2001, ApJ, 551, 807

Draine, B.T., & Li, A., 2007, ApJ, 657, 810

Galliano, F., Madden, S.C., Tielens, A.G.G.M., Peeters, E., & Jones, A.P., 2008, ApJ,679, 310

Gordon, K.D., Clayton, G.C., Misselt, K.A., Landolt, A.U., & Wolff, M.J., 2003, ApJ,594, 279

Hoang, T., Draine, B.T., & Lazarian, A., 2010, ApJ, 715, 1462

Joblin, C., Leger, A., & Martin, P., 1992, ApJ, 393, L79

Kogut, A., Banday, A.J., Bennett, C.L., et al., 1996, ApJ, 464, L5

Lazarian, A., & Draine, B.T., 2000, ApJ, 536, L15

Leger, A., D’Hendecourt, L., & Defourneau, D., 1989, A&A, 216, 148

Li, A., & Draine, B.T., 2001a, ApJ, 554, 778

Li, A., & Draine, B.T., 2001b, ApJ, 550, L213

Li, A., & Draine, B.T., 2002, ApJ, 576, 762

Mason, B.S., Robishaw, T., Heiles, C., Finkbeiner, D., & Dickinson, C., 2009, ApJ, 697,1187


Mathis, J.S., Mezger, P.G., & Panagia, N., 1983, A&A, 128, 212

Munoz-Mateos, J.C., Gil de Paz, A., Boissier, S., et al., 2009, ApJ, 701, 1965

Sandstrom, K.M., Bolatto, A.D., Draine, B.T., Bot, C., & Stanimirovic, S., 2010, ApJ,715, 701

Silsbee, K., Ali-Haimoud, Y., & Hirata, C.M., 2010 [arXiv1003.4732]



Voit, G.M., 1992, MNRAS, 258, 841

Weingartner, J.C., & Draine, B.T., 2001, ApJ, 548, 296

Ysard, N., Miville-Deschenes, M.A., & Verstraete, L., 2010, A&A, 509, L1

Ysard, N., & Verstraete, L., 2010, A&A, 509, A12

Zubko, V., Dwek, E., & Arendt, R.G., 2004, ApJS, 152, 211


DIALECTICS OF THE PAH ABUNDANCE TREND WITHMETALLICITY

F. Galliano1

Abstract. This paper reviews the various processes that have been pro-posed to explain the observed trend of PAH strength with metallicity.It summarizes a study showing that, although PAH destruction by hardradiation is important in low-metallicity environments, it is not suffi-cient to explain their paucity. In these systems, the deficiency of theirformation mechanism, related to stellar evolution, has to be invoked.

1 Introduction: An intricate combination of processes

From an extragalactic point of view, the strength of the aromatic features appearsto be correlated with the metallicity of the environment. The general trend hasbeen discussed by many authors. It was first demonstrated by Madden et al. (2006)using ISO spectra, and by Engelbracht et al. (2005) with Spitzer broadbands.

At the scale of a galaxy, the aromatic-feature-to-mid-IR-continuum intensityratio appears to be a monotonic rising function of the metallicity of the gas, Zgas.This relation contains a lot of scatter, and the metallicity is uncertain and some-times difficult to define due to its inherent gradient within large galaxies. How-ever, there is no reason to a priori consider that this trend defines two regimesabove and below 12 + log(O/H) � 8.0. The notion of such a threshold was intro-duced by Engelbracht et al. (2005), when studying the IRAC8 μm/MIPS24 μm ratio.Galliano et al. (2008, Figure 1a of the present paper) showed that this was a biasdue to the fact that, for low PAH-to-continuum ratios, the IRAC8 μm band stopsbeing a good tracer of the PAH strength, as it becomes dominated by the con-tinuum intensity. With such a threshold, several galaxies, which are clearly PAHdeficient, would fall in the regime where “normal” galaxies are.

Several scenarios have been proposed to explain the origin of the general trend.

1. PAHs are known to be massively destroyed in regions bathed with hard UVphotons. The filling factor of molecular clouds in dwarf galaxies is lower

1 Service d’Astrophysique – Laboratoire AIM, CEA/Saclay, L’Orme des Merisiers,91191 Gif-sur-Yvette, France; e-mail: [email protected]



Fig. 1. (a) Ratio of the IRAC8 μm and MIPS24 μm broadbands of our sample,

as a function of the observed metallicity. For low PAH-to-continuum ratios, the

IRAC8 μm/MIPS24 μm saturates. (b) Modelled contribution of H ii regions to the IR lu-

minosity, as a function of the observed metallicity. (c) Modelled galaxy age (from stellar

population synthesis), as a function of the observed metallicity. (d) Modelled PAH-to-

dust mass ratio as a function of the observed metallicity. Those 4 panels all come from

Galliano et al. (2008). On each panel, the grey stripe shows the linear correlation to the

data points ±1σ.

than in normal metallicity galaxies (Madden et al. 1997). The PAHs couldsimply be destroyed on large scales in low-metallicity environments, due tothe lower opacity of their dust depleted interstellar medium (ISM). Thisexplanation was supported by the anticorrelation of the PAH-to-continuumintensity ratios with [Ne iii]15.56 μm/[Ne ii]12.81 μm (Madden et al. 2006). Thelatter line ratio is a tracer of the very young, ionizing stellar populations.

F. Galliano: PAH abundance with metallicity 45

2. Alternatively, O’Halloran et al. (2006) have proposed that PAHs could bedestroyed by the numerous supernova (SN) shock waves sweeping the ISMof dwarf galaxies. However, observations of Galactic supernova remnantsdemonstrate that the blast waves destroy all dust species (PAHs and theother grains; e.g. Reach et al. 2002) and therefore they do not pose a straight-forward explanation for the selective destruction of the PAHs.

3. Finally, Dwek (2005) noted that the contribution to the enrichment of theISM by the stellar progenitors commonly associated to PAH production, theAGB stars, was also a rising function of metallicity. Therefore the trend dis-cussed in this paper could originate in the fact that PAHs are less abundantlyproduced in chemically young systems.

Correlation is not causality. Since the different mechanisms invoked above are afunction of metallicity, it is difficult to find which one is at the origin of the trend.Looking at only one of these effects independently will always bias the conclusion.To really tackle the origin of this trend, it is needed to confront these variouspossible processes, and quantify their respective contributions.

2 An estimate of the PAH content of galaxies

To address this issue, we have studied the competition of these various effectswithin a sample of 35 nearby galaxies (1/50 Z� � Zgas � 2 Z�). We mod-elled the global UV-to-radio spectral energy distribution (SED) of each galaxy, inorder to derive its PAH and dust-to-gas mass ratios. This modelling was doneself-consistently, taking into account realistic dust properties, consistent stellarevolution and PAH destruction in H ii regions. The detail of this modeling ispresented by Galliano et al. (2008).

The contribution of H ii regions to the IR luminosity is shown in Figure 1b.This contribution is constrained by the observations of the radio free-free and mid-IR thermal continua. We assume that PAHs are fully depleted in H ii regions. Ittherefore quantifies the effect of PAH destruction on the total SED. This effect isimportant when considering its impact on the IRAC8 μm/MIPS24 μm ratio, but isvery limited on the IRAC8 μm/MIPS160 μm which is the main constraint on theaverage PAH mass fraction.

Figure 1c shows the galaxy age, derived from stellar population synthesis andfit to the near-IR, as a function of metallicity. It demonstrates that our derivedstar formation history is consistent with the independent metallicity estimate ofthe system. This stellar population modelling is important for the subsequentchemical evolution modelling.

Finally, Figure 1d shows the evolution with metallicity of our derived PAH-to-dust mass ratio. Although there is a lot of scatter (partly due to the propagationof the observational uncertainties through the model), this panel shows a clear,relatively smooth, increase with metallicity of the PAH mass fraction.


3 A global point view on PAH evolution in galaxies

To interpret the trend in Figure 1d, we developed a dust evolution model, thattakes into account the metal and dust enrichment of the ISM by stars (Gallianoet al. 2008). Assuming a continuous star formation history (following a Schmidt-Kennicutt law), the model predicts at each time the injection rate of the variouselements and dust species. In particular, it computes the evolution of the carbonand silicate dust as a function of metallicity. We independently track the dustproduction associated with massive and AGB stars. We consider the destructionrate of dust in the ISM by SN blast waves to be proportional to the SN rate.

Figure 2 compares the observed trends of PAH and dust-to-gas mass ratios,with metallicity, to the theoretical evolution of carbon dust produced by AGB starsand dust produced by massive stars. The PAH evolution is strikingly coincidentwith the production of carbon dust by low-mass stars (see Cherchneff in thisvolume, for a discussion on PAH condensation in AGB stars). This comparisontherefore suggests that the main origin of the PAH trend with metallicity is aresult of the delayed injection of these molecules by AGB stars. Indeed, AGBstars start enriching the ISM after their death, �400 Myr after the beginning ofthe star formation process, when the galaxy has already been enriched by massivestars.

The stellar origin of dust is widely debated (Draine 2009, for a review). Dustis heavily processed in the ISM and reforms in dense clouds. A part of the ob-served ISM dust content probably does not have a stellar origin. Although ourmodel accounts for the production of dust by various stellar progenitors, thereis no assumption made on the actual location and mechanism of condensation.The comparison of our theoretical dust evolution trend to the PAH abundancein galaxies (Fig. 2) would not suffer if the actual dust condensation took placea few 107 yr after the metal injection in the ISM. Consequently, our scenario isnot in contradiction with dust reformation in the ISM, as long as it occurs onshort timescales after the element injection. Dust precursors formed in the stellarenvelopes could be injected in the ISM. Although their mass could account foronly a few percent of the ISM budget, their presence in the ISM could be crucialto accrete more material and grow the observed ISM grains. That would explainwhy the dust abundance in galaxies would be closely related to the injection ofthose grain seeds.

Another concern on these trends is the fact that many dwarf galaxies harbouran old stellar population. It is an indication that these objects had a complexstar formation history, and that they are older than indicated by their metallicity.However, the contribution of this old stellar population to the enrichment of theISM is not significant, otherwise their metallicity would be higher. Moreover, thisstellar population does not dominate the integrated SED of these objects (Fig. 1b).Finally, Lee et al. (2009) studied a sample of local dwarf galaxies and showed thatthe star formation bursts were only responsible for about a quarter of the totalstar formation in the overall population.

F. Galliano: PAH abundance with metallicity 47

Fig. 2. Comparison between our dust evolution model and the derived trend of PAH

and dust-to-gas mass ratios (ZPAH = MPAH/Mgas and Zdust = Mdust/Mgas) with the

observed gas metallicity. The grey envelopes show the result of our dust evolution model.

The various curves within the envelopes represent the spread caused by assuming different

star formation rates and different dust destruction efficiencies by SN blast waves. No

parameter has been adjusted to fit the dust evolution trends to the observed dust content.

4 Additional confirmations from spatially resolved studies

Although our global approach is relatively crude, because of its lack of spatialresolution, it provides general trends of the various processes and allows a com-parison of their respective contributions. Several studies have addressed some ofthese issues by looking at spatially resolved observations.


Munoz-Mateos et al. (2009) by looking at radial trends within spiral galaxiesconfirmed the relation between PAH and the evolution of AGB stars. They evenfound indications of the reversing of the trend at high metallicity, explained by thefact that the enrichment becomes dominated by O-rich low-mass stars (Fig. 2).

In the Large Magellanic Cloud (LMC), the spatial distribution of the PAHmass fraction is associated with the stellar bar (Paradis et al. 2009). It has beenindependently confirmed by Meixner et al. (2010). However, this is not the case inthe Small Magellanic Cloud (SMC; � 1/6 Z�; Sandstrom et al. 2010, also in thisvolume). In the SMC, the PAHs appear to be massively depleted in the diffuseISM, but they are abundant in PDRs, although their abundance is significantlylower than in the Milky Way (�0.2−0.4 Galactic PAH abundance). Consequently,this study shows, with spatial resolution, that despite the photodestruction effectsdominate in the diffuse ISM, the PAHs are underabundant even in regions wherethey are shielded from the hard radiation field. It therefore supports the idea thatthe photodestruction processes are not sufficient to account for the trend of PAHabundance with metallicity, and that the deficiency of their production has to beinvoked.

References

Draine, B.T., 2009, ed. in T. Henning, E. Grun, & J. Steinacker, Astron. Soc. Pac. Conf.Ser., 414, 453

Dwek, E., 2005, ed. in C.C. Popescu, & R.J. Tuffs, AIP Conf. Proc. 761: The SpectralEnergy Distributions of Gas-Rich Galaxies, 103

Engelbracht, C.W., Gordon, K.D., Rieke, G.H., et al., 2005, ApJ, 628, L29

Galliano, F., Dwek, E., & Chanial, P., 2008, ApJ, 672, 214

Lee, J.C., Kennicutt, R.C., Jose G. Funes, S.J., et al., 2009, ApJ, 692, 1305

Madden, S.C., Galliano, F., Jones, A.P., & Sauvage, M., 2006, A&A, 446, 877

Madden, S.C., Poglitsch, A., Geis, N., et al., 1997, ApJ, 483, 200

Meixner, M., Galliano, F., Hony, S., et al., 2010, A&A, 518, L71

Munoz-Mateos, J.C., Gil de Paz, A., Boissier, S., et al., 2009, ApJ, 701, 1965

O’Halloran, B., Satyapal, S., & Dudik, R.P., 2006, ApJ, 641, 795

Paradis, D., Reach, W.T., Bernard, J., et al., 2009, AJ, 138, 196

Reach, W.T., Rho, J., Jarrett, T.H., & Lagage, P.-O., 2002, ApJ, 564, 302

Sandstrom, K.M., Bolatto, A.D., Draine, B.T., et al., 2010, ApJ, 715, 701


THE SHAPE OF MID-IR PAH BANDS IN THE UNIVERSE

O. Berne1, P. Pilleri2, 3 and C. Joblin2,3

Abstract. A large number of galactic and extragalactic sources ex-hibit a mid-IR spectrum that is dominated by PAH emission. Look-ing at these spectra in more details reveals a strong variability in theshape/position of the observed features, depending on the observedsource, or even the region within a source. In this article, we presentthe results of an analysis that has allowed us to decompose these spec-tra into components having a physical meaning. Most, if not all PAHdominated mid-IR spectra of HII regions, PDRs, protoplanetary disks,galaxies etc. can be fitted efficiently using a combination of these com-ponents. The results of these fits provide further insight in the compo-sition of the emitting material and the local physical conditions. In theframe of the future IR space missions (JWST, SPICA), this approachcan be very useful to probe the physical conditions in distant galaxies.

1 Introduction

The ubiquitous mid-IR emission bands, widely observed in the spectra of dustyastrophysical sources (from protoplanetary disks to starburst galaxies), are com-monly attributed to the emission of polycyclic aromatic hydrocarbons (PAHs).However, because these bands are due to nearest neighbor vibrations of the C-Cor C-H bonds, they are not specific to individual PAH species. Because essen-tially any PAH molecule will carry these bonds, the emission bands associated tovibrations are not discriminant to specific PAH molecules, thus preventing fromany individual identification. However, these bands do carry some information, inparticular their position in wavelength and their relative intensity are known tovary strongly depending on the chemical evolution of their carriers that is driven

1 Leiden Observatory, Leiden University, Niels Bohrweg 2, 2333 CA Leiden,The Netherlands2 Universite de Toulouse, UPS, CESR, 9 avenue du Colonel Roche, 31028 Toulouse Cedex 4,France3 CNRS, UMR 5187, 31028 Toulouse, France



0

0.1

0.2

λ (∝m)

0

0.05

0.1

0.15

0.2

λ (∝m)

I ν (A

.U.)

6 7 8 9 10 11 12 13 14 15 160

0.05

0.1

0.15

0.2

0.25

0.3

λ (μm)

eVSGs

Neutrals

Cations

57.6 50.4 43.2 21:01:36.0 28.8 21.6 14.4

30.0

11:00.0

30.0

68:10:00.0

30.0

09:00.0

30.0

08:00.0

RA (2000)

Dec

(20

00)

Fig. 1. Left : extracted spectra attributed to evaporating VSGs, PAH0 and PAH+ in

NGC 7023. Right : associated distribution maps of the three populations: eVSGs in red,

PAH0 in green and PAH+ in blue. The white circle shows the position of the exciting

star. Colors combine as in an RGB image i.e. green (PAH0)+red (eVSGs)=yellow.

by physical conditions. This has been well demonstrated on the basis of ISO obser-vations (see e.g. Peeters et al. 2002; Rapacioli et al. 2005) and more recently withSpitzer (see e.g. Werner et al. 2004; Berne et al. 2007). The observed IR spectrumusually consists of a set of bands that are most prominent at 6.2, 7.7, 8.6, 11.3, and12.7 μm. It was established that the modification of the shape of this spectrumcan be attributed to the chemical evolution of the emitting populations (Peeterset al. 2002; Hony et al. 2001) as the local physical conditions change. In partic-ular, models (Tielens 2005) and observations (Joblin et al. 1996; Galliano et al.2008) have shown that the variations of the 6.2 (or 8.6) to 11.3 μm band intensityratio (I6.2/I11.3) evolves with the ionization parameter γ = G0 × √

T/nH whereG0 is the intensity of the UV radiation field in Habing units, T is the gas tempera-ture and nH the total hydrogen nucleus density. Following this work, Berne et al.(2009b) have shown that the combination of the measurement of I6.2/I11.3 and ofthe ratio between the H2 0-0 S(3) and S(2) line intensities, respectively at 9.7 and12.3 μm, allows to derive the individual values of T , G0 and nH when they fall inthe ranges T = 250 − 1500 K, nH = 104 − 106 cm−3, G0 = 103 − 105 respectively.

2 Identification of underlying spectral components

Rapacioli et al. (2005) and Berne et al. (2007) extracted the “pure” spectra ofthree different populations of PAH-like grains, namely: neutral PAHs (PAH0), ion-ized PAHs (PAH+) and evaporating very small grains (eVSGs, see

O. Berne et al.: The Shape of Mid-IR PAH Bands in the Universe 51

6 7 8 9 10 11 12 130

0.5

1

1.5

6 7 8 9 10 11 12 130

0.5

1

1.5

6 7 8 9 10 11 12 130

0.5

1

1.5

Inte

nsity

(A

rbitr

ary

Uni

ts)

6 7 8 9 10 11 12 130

0.5

1

1.5

t

PAH0

PAH+

eVSG

PAHx

Spectra adapted from BSS results

Spectra of large (>100 carbon atoms) ionized PAHs inspired from

Quantum Chemistry results [Joblin et al 2008, Bauschlischer 2008]

Fig. 2. Template spectra based on the results of Berne et al. (2007) and Rapacioli et al.

(2005). In addition, the PAHx spectrum introduced by Joblin et al. (2008).

Pilleri et al. 2010) in the reflection nebula NGC 7023 using blind signal sepa-ration methods. This is interpreted as evidence that the evolution of the shapeof the mid-IR bands is related to the chemical evolution under the effect of UVphotons (see Fig. 1): free PAH molecules are formed from photoevaporation ofvery small grains (see Rapacioli et al. in this volume for a review of possible can-didates for eVSGs). The main evolution observed for the PAH population is toget ionized closer to the star. Based on these previous results, Joblin et al. (2008)and Berne et al. (2009a), were then able to construct a simple model comprisingof eVSG, PAH0 and PAH+ mid-IR template spectra (Fig. 2.). While constructingthis model, they found that an additional population of PAHs consisting of large(>100 atoms of carbon) and ionized PAHs (called PAHx) had to be introduced inorder to reproduce the observed spectrum of highly UV-irradiated environments.Other authors (Bauschlisher 2008; Tielens 2008; Geers 2008) have also pointedout the existence of this population of large PAHs.

3 PAHs/eVSGs: Tracers of physical conditions

The templates defined in the previous section can be used to fit observed mid-IRspectra of different astrophysical objects such as planetary nebulae (Joblin et al.2008), protoplanetary disks (Berne et al. 2009a), and galaxies (Vega et al. 2010).One can then relate the results of the fits to the physical conditions as probed byindependent tracers, in order to evidence empirical connections. This was done e.g.by Berne et al. (2009a) for protoplanetary disks (Fig. 3). The authors have shownthat the fraction of emission attributed to eVSGs in the mid-IR spectra of a sampleof 12 disks correlates well with decreasing stellar temperature (Fig. 3). This wasinterpreted as the more efficient destruction of eVSGs in UV-rich environments.


Fig. 3. Left : example of fits (gray), using the template spectra of Figure 2 (corrected for

extinction on line of sight), to the observed mid-IR spectra of two protoplanetary disks.

Right : the fraction of “evaporating VSG” emission (%) that was deduced from the fits

for a set of 12 observed spectra, as a function of the UV luminosity of the central star

(i.e. spectral type).

Recently, a detailed and quantitative analysis comparing the results of PDR modelsto the results obtained by fitting the set template spectra to mid-IR spectral cubesobtained with Spitzer has shown that, indeed, the eVSG abundance can be usedas a tracer of the intensity of the UV field (Pilleri et al. 2010). The templatefitting approach can be applied to the mid-IR spectral cubes of nearby galaxies,which in the end provides the spatial distribution of the different populations ofPAHs/eVSGs (Fig. 4). This spatial distribution then informs on the local physicalconditions of the ISM in different parts of the galaxy. Here, in the case of M 82,the outflow is PAH0-rich, implying the presence of large amounts of molecular gas(as also suggested by Micelotta et al. 2010), shielding the PAHs from ionizationand allowing quick recombination of PAH+ with electrons. The disk, on the otherhand, is filled with PAH+/x consistent with the intense star-formation activityproducing strong UV fields. Recently, Vega et al. (2010) have applied this fittingapproach to the Spitzer spectra of spatially unresolved early type galaxies, andshown the pristine nature of their carbonaceous dust content, implying a constantreplenishment of dust by carbon stars. Eventually, one could also use this fittingtechnique to interpret the mid-IR spectra of galaxies at high redshift and derivetheir properties.

O. Berne et al.: The Shape of Mid-IR PAH Bands in the Universe 53

Fig. 4. Example of fit (blue line), using the template spectra of Figure 2 (corrected for

extinction on the line of sight), to an observed mid-IR spectrum taken in the cube of

M82 in the outflow region. Achieving the fit on all the spatial positions of the spectral

cube, the spatial distribution map of the PAH/eVSG populations can be built (inset).

In the inset, white contours show the 3.6 μm -mostly- stellar emission.

1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 20100

50

100

150

200

250

year

PA

H p

aper

s

1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 20100

10

20

30

40

50

60

70

80

90

100

year

% o

f Xga

l PA

Hs

Fig. 5. Left : number of published papers each year regarding PAHs in astronomy in the

last 20 years. Right : percentage of papers regarding PAHs that focus on PAHs observed

in external galaxies in the last 20 years. Source: NASA-ADS.

4 Perspectives in astronomical PAH science

The improvement of ground- and space-based instruments in the mid-IR haveprovided us with an incredible legacy of astronomical PAH spectra. This has, inthe last 20 years, motivated each year more studies (Fig. 5). As we have brieflyillustrated in the previous section, the PAH spectrum contains a wealth of infor-mation on the physical conditions of the emitting source. More and more, thesebands are used as a tracer of physical conditions in UV-irradiated astrophysicalenvironments (protoplanetary disks, galaxies, . . . ) because they are easy to detect(Fig. 5). This is why knowing the detailed composition of PAHs, their evolutionand chemistry in space remains one of the key challenges of modern astrophysics.The ISO and Spitzer legacies, the Herschel surveys now (see Joblin et al. in this


volume), and the SPICA (Berne et al. 2009c), JWST and ELT observations inthe future, will contribute to this task.

References

Berne, O., Joblin, C., Mulas, G., Tielens, A.G.G.M., & Goicoechea, J.R., 2009c, SPICAjoint European/Japanese Workshop, July, 2009

Berne, O., Fuente, A., Goicoechea, J.R., et al., 2009b, ApJL, 706, 160

Berne, O., Joblin, C., Fuente, A., & Menard, F., 2009a, A&A, 495, 827




Joblin, C., Szczerba, R., Berne, O., Szyszka, C., et al., 2008, A&A, 490, 189

Joblin, C., Tielens, A.G.G.M., Geballe, T.R., & Wooden, D.H. 1996, ApJ, 460, L119



Micelotta, E.R., Jones, A.P., & Tielens, A.G.G.M., 2010, A&A, 510, 37

Tielens, 2008, ARA&A, 46, 289

Tielens, A., 2005 (Cambridge University Press)

Vega, O., Bressan, A., Panuzzo, P., et al., 2010, ApJ, 721, 1090

Werner, M.W., Uchida, K.I., Sellgren, K., et al., 2004, ApJS, 154, 309


AKARI NEAR-INFRARED SPECTROSCOPY OF 3 MICRONPAH AND 4 MICRON PAD FEATURES

T. Onaka1, I. Sakon1, R. Ohsawa1, T. Shimonishi1, Y. Okada2,M. Tanaka3 and H. Kaneda4

Abstract. Near-infrared (NIR; 2.5–5 μm) low-resolution (λ/Δλ ∼ 100)spectra were obtained for a number of Galactic and extragalactic ob-jects with the Infrared Camera (IRC) in the AKARI warm mission.These data provide us with the first opportunity to make a system-atic study of the 3.3–3.5 μm PAH features in a galactic scale as wellas within an object. Whereas the 3.3 μm band is well resolved in mostspectra, the 3.5 μm band is not clearly separated from the 3.4 μm bandin the IRC spectrum. The intensity ratio of the summation of the 3.4and 3.5 μm bands to the 3.3 μm band shows a tendency to increasetowards the Galactic center, although a large variation in the ratio isalso seen in a local scale. A search for deuterated PAH features in the4μm region is carried out in IRC NIR spectra. Emission lines originat-ing from the ionized gas together with the detector anomaly hamperan accurate search at certain wavelengths, but little convincing evi-dence has so far been obtained for the presence of significant featuresin 4.2–4.7 μm. A conservative upper limit of a few percents is obtainedfor the integrated intensity ratio of the 4.4–4.7 μm possible features tothe 3.3–3.5 μm PAH features in the spectra so far obtained.

1 Introduction

The presence of infrared band features (hereafter PAH bands) in the diffuse Galac-tic radiation, which are attributed to polycyclic aromatic hydrocarbons (PAHs)or PAH-related materials, has been first confirmed for the 3.3μm band by theAROME balloon experiment (Giard et al. 1988) followed by observations with

1 Department of Astronomy, Graduate School of Science, University of Tokyo, Tokyo 113-0033, Japan; e-mail: [email protected] I. Physikalisches Institut, Universitat zu Koln, 50937 Koln, Germany3 Center for Computational Sciences, University of Tsukuba, Ibaraki 305-8577, Japan4 Graduate School of Science, Nagoya University, Aichi 464-8602, Japan



the Infrared Telescope in Space (IRTS; Tanaka et al. 1996). Detection of mid-infrared PAH bands in the interstellar medium (ISM) was made simultaneouslywith IRTS (Onaka et al. 1996) and Infrared Space Observatory (ISO; Mattilaet al. 1996). Since then a number of investigations have been carried out for thevariations in the MIR PAH bands in Galactic and extragalactic objects includingthe diffuse Galactic radiation (e.g., Chan et al. 2001; Peeters et al. 2002; Sakonet al. 2004; Rapacioli et al. 2005; Berne et al. 2007; Smith et al. 2007) based onIRTS, ISO and Spitzer observations. However no systematic study has been madefor the 3 μm PAH band features in a galactic scale except for the studies of theband variation among a small number of objects and within an object by usingground-based telescopes (e.g., Geballe et al. 1989; Joblin et al. 1996; Song et al.2003) because of a scarcity of sensitive spectrometers in the 3 μm region.

AKARI, the Japanese satellite mission dedicated to infrared astronomy(Murakami et al. 2007), used up 180 liter liquid Helium in 2007 August, 550 daysafter the launch, and completed its cold mission. However, the on-board cryocoolerstill keeps the telescope and instrument low enough to continue near-infrared (NIR)observations with the Infrared Camera (IRC; Onaka et al. 2007). The IRC hasa low-resolution spectroscopic capability (Ohyama et al. 2007) in addition to theimaging, which enables us to carry out sensitive spectroscopic observations in theNIR. In this paper we present the latest results of a study of the 3 μm PAH bandsbased on AKARI/IRC spectroscopic observations carried out during the AKARIwarm mission (Onaka et al. 2010) together with the results of a search for featuresof deuterated PAHs in the 4 μm region.

2 IRC spectroscopy and results

The IRC spectroscopy has the slit and slit-less modes with a choice of the dis-persers: prism (NP) and grism (NG) (see Ohyama et al. 2007 for details). Forslit-less point source spectroscopy with the grism, a small window is used to avoidoverlapping with other sources. A number of various types of objects, includingyoung stellar objects, planetary nebulae, and external galaxies, have so far beenobserved in this mode in the AKARI warm mission (e.g., Shimonishi et al. 2010).

The IRC slit spectroscopy has a choice of the slit: medium (Ns) and narrow(Nh). The spectral resolution with Ns is about 100, whereas Nh provides about150. Both are long slits (∼0.′8) and spatial information can be extracted along theslit. The data used in the present study were taken in the Ns slit spectroscopy. Theobservations were carried out as part of the AKARI mission program InterstellarMedium in our Galaxy and nearby galaxies (ISMGN; Kaneda et al. 2009). Morethan 100 Galactic objects, most of which are HII-PDR complexes, and about 100positions on the Galactic plane (diffuse Galactic radiation) have been observed inthis program, part of which are used in the present investigation.

Figure 1 shows examples of the spectra of the Galactic plane. They clearlydetect the 3.3, 3.4, and 3.5μm PAH features together with hydrogen recombinationlines originating from the ionized gas (e.g., Brα at 4.05μm and Brβ at 2.63μm).The 3.3μm band is well resolved, but the separation of the 3.4 and 3.5μm bands

T. Onaka et al.: AKARI NIR Spectroscopy of PAH and PAD 57

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

�

�� μ!�

�� "�� "�#$ �� "��#$

�� μ!� �� μ!�

Fig. 1. Examples of AKARI/IRC NIR spectra of the Galactic plane: (a) l ∼ 0◦, (b) l ∼78◦, and (c) l ∼ 178◦.

is not always clear. The spectra also indicate a significant variation in the PAHfeatures. The spectrum of the Galactic center region (a) shows the absorptionfeatures of H2O and CO2 ices at 3.0 and 4.3μm, respectively, as well as gaseousCO absorption at 4.6μm and Brα emission at 4.05μm.

To derive the band intensity, we first make a spline fit with the continuum andthen fit the PAH features in the 3μm region. A combination of two Lorentzians(for the 3.3 and 3.4μm bands) and one Gaussian (for the 3.5μm band) providesthe best fit; however, this does not indicate the nature of the bands because the3.5μm band consists of more than one components, which are not resolved with thepresent resolution. Since the separation of the 3.4 and 3.5μm bands is not alwaysclear, only a summation of the two band strengths is discussed in the following.

Figure 2a shows the intensity ratio of the summation of the 3.4 and 3.5μmbands to the 3.3μm band against that of the 3.4 to 3.3μm band. As describedabove, the horizontal axis may not be well defined and is used here only for theillustration purpose. The compact sources (mostly HII-PDR complexes) have theratio in a relatively narrow range (0.2–0.5), while the ratio of the diffuse Galacticradiation shows a large variation. Figure 2b plots the intensity ratio in the diffuseradiation against the Galactic longitude. It suggests a global trend that the ratioincreases towards the Galactic center region, though the scatter is large. Thelatter is compatible with the variations seen within an object and among objectsin previous studies (e.g., Geballe et al. 1989; Joblin et al. 1996).

3 Search for deuterated PAH features

The abundance of deuterium in the ISM is an important parameter in the studyof the galaxy evolution since deuterium was formed at the beginning of the Uni-verse and has been gradually destroyed in the stellar interior (“astration”). Itsabundance must be strongly linked to the cosmological parameters and the chemi-cal evolution of the Galaxy. Recent far-ultraviolet observations, however, indicatethat the interstellar deuterium abundance is significantly lower than model pre-dictions. In addition they suggest several pieces of evidence that deuterium isdepleted onto dust grains (Linsky et al. 2006). Draine (2006) proposed that


��

��

��

��%�

��$�

��

��

��

&��μ!�'&��μ!�� % ��$ �� $� �� %��

��(��

)&��

�μ!

�*&

��

μ!�+

�

'&��

��μ!

�

��

�,!-��,��

(� ��,��

�� .��$�� /��$�

Fig. 2. (a) The intensity ratio of the summation of the 3.4 and 3.5 μm bands to the

3.3 μm band against that of the 3.4 to 3.3 μm band. The open squares indicate “compact”

sources, which are mostly HII region-PDR complexes (see text), while the crosses show

the diffuse radiation from the Galactic plane. (b) The Galactic longitude dependence of

the intensity ratio of the summation of the 3.4 and 3.5 μm bands to the 3.3 μm band for

the diffuse Galactic radiation. The crosses indicate those of 0 ≤ l < 180◦ and the open

squares show those of 180 ≤ l < 360◦.

deuterium may be contained in interstellar PAHs since hydrogen in PAHs couldbe easily replaced by deuterium in low temperature environments. The 3.3 and3.4μm PAH bands should be shifted to the 4.3–4.7μm region once they are deuter-ated. If the observed depletion can be attributed solely to deuteration of PAHs,then the deuterium fraction in PAHs can be as large as 0.3. Peeters et al. (2004)reported a possible detection of the 4.4 and 4.65μm features in the Orion bar andM 17 based on ISO/SWS spectra. They derived the integrated intensity ratio ofthe deuterated PAH features to the 3.3 and 3.4μm PAH features to be 0.17±0.03and 0.36 ± 0.08 for the Orion bar and M 17, respectively. However, the detectionwas marginal even for the best case (4.4σ) and needs to be confirmed by fur-ther observations. The IRC spectroscopy offers the best opportunity to search fordeuterated PAH features in the 4μm, since the predicted spectral range is stronglyblocked by the terrestrial atmosphere.

A couple of problems have to be taken into account before making a reliablesearch for features in the 4 μm region in IRC spectra. First, the IRC NIR detectorarray shows fixed patterns in the direction vertical to the dispersion, which couldproduce spurious features in the spectrum. Appreciable features are in fact seenoccasionally at around 3.8, 3.95, and 4.4μm. The last one unfortunately coincideswith the wavelength of one of the predicted deuterated PAH features. For theslit-less spectroscopy this does not make a problem since the spurious features canbe removed by dithering operations, but it could be a serious problem for the slitspectroscopy. A spectral flat is created for the slit spectroscopy from the spectraof the central part of M 31 by assuming that it is smooth without any appreciablefeatures, and then applied to the observed spectra.

T. Onaka et al.: AKARI NIR Spectroscopy of PAH and PAD 59

0��,��

1��!,(�

��

��

��

��

��

��

��

��

��

��

��

��

��

�

��

��

%��

2��

��

��

�� μ!� �� μ!�

Fig. 3. (a) AKARI/IRC spectrum of M17. The solid line shows the observed spectrum

corrected for the detector anomaly, whereas the dotted line indicates the spectrum with

hydrogen recombination lines that are calculated from Brα based on the case B con-

ditions (see text). Note that the HeI line at 4.3 μm is not included in the calculation.

(b) AKARI/IRC spectrum of a region without recombination lines.

The other issue is the evaluation of the contribution of emission lines from theionized gas. In particular, Pfβ at 4.65μm appears at a wavelength near a predictedfeature of deuterated PAHs. Contribution from the ionized gas, if any, must beremoved from the spectrum. Assuming the case B conditions with the electrontemperature of 7500 K and the density of 104 cm−3, the line intensity of Pfβ isestimated from Brα. The intensities of Brβ and Pfγ are also used to estimatepossible extinction.

Figure 3a shows an IRC spectrum (solid line) of M 17 together with the spec-trum with the predicted line emission (dotted line). The model well accountsfor the observed hydrogen emission lines. The observed 4.65μm line seems to beslightly stronger than the predicted Pfβ. There is also excess emission around4.3μm. Part of the excess can be attributed to the He recombination line at4.3μm, which is not included in the line calculation. Even taking account of pos-sible features remaining around 4.4 and 4.65μm, however, the summation of theintegrated intensities of these features amounts only 3% of the summation of theintegrated intensities of the 3μm PAH features.

Spectra of the regions that do not have appreciable contribution from theionized gas have a less uncertainty in the removal of Pfβ and could impose a morestringent constraint on the possible feature intensities. Figure 3b shows one of theexamples, which was obtained towards the Galactic plane of l ∼ 10◦. The spectralflat has been applied. It does not show any significant features in the 4μm region,while the 3.3–3.5μm features are clearly seen. A conservative upper limit for theratio of the integrated intensities is obtained as 2%.

4 Summary

The IRC NIR spectroscopy in the AKARI warm mission provides the first oppor-tunity to make a systematic study of the 3 μm PAH features in various objects.


The intensity ratio of the 3.4 and 3.5μm bands to the 3.3μm band is relativelyconstant for HII-PDR complexes. The intensity ratio of the diffuse Galactic emis-sion, on the other hand, shows a large variation in a galactic scale, suggestinga tendency that the ratio increases towards the Galactic center region, althoughthe scatter is large. A search for deuterated PAH features in the 4μm region ofthe IRC spectra of Galactic objects gives a conservative upper limit of 2–3% forthe integrated intensity of the deuterated PAH features relative to the 3.3–3.5μmfeatures. It should be noted that this ratio cannot directly be converted into theabundance ratio because the oscillator strength of isotopic species in the vibrationtransition could be different. Moreover the difference in the excitation conditionsbetween the 3 and 4μm regions must also be taken into account properly.

This work is based on observations with AKARI, a JAXA project with the participation of ESA.The author thanks all the members of the AKARI project for their continuous support. Thiswork is supported by a Grant-in-Aid for Scientific Research and a Grant-in-Aid for challengingExploratory Research from the JSPS.

References


Chan, K.-W., Roellig, T.L., Onaka, T., et al., 2001, ApJ, 231, 183

Draine, B.T., 2006, ASP Conf. Ser., 348, 58

Joblin, C., Tielens, A.G.G.M., Geballe, T.R., & Wooden, D.H., 1996, ApJ, 458, 610

Geballe, T.R., Tielens, A.G.G.M., Allamandola, L.J., Moorhouse, A., & Brand, P.W.J.L.,1989, ApJ, 341, 278

Giard, M., Serra, G., Caux, E., Pajot, F., & Lamarre, J.M., 1988, A&A, 201, L1

Kaneda, H., Koo, B.-C., Onaka, T., & Takahashi, H., 2009, Adv. Sp. Res., 44, 1038

Linsky, J., Draine, B.T., Moos, H.W., et al., 2006, ApJ, 647, 1106

Mattila, K., Haikala, L.K., Laureijs, R., et al., 1996, A&A, 315, L353

Murakami, H., Baba, H., Barthel, P., et al., 2007, PASJ, 59, S309

Ohyama, Y., Onaka, T., Matsuhara, H., et al., 2007, PASJ, 59, S411

Onaka, T., Yamamura, I., Tanabe, T., Roellig, T.L., & Yuen, L., 1996, PASJ, 48, L59

Onaka, T., Matsuhara, H., Wada, T., et al., 2007, PASJ, 59, S401

Onaka, T., Matsuhara, H., Wada, T., et al., 2010, Proc. SPIE, 7731, 77310M

Peeters, E., Hony, S., Van Kerchhoven, C., et al., 2002, A&A, 390, 1089

Peeters, E., Allamandola, L.J., Bauschlicher, C.W., Jr., et al., 2004, ApJ, 604, 252


Sakon, I., Onaka, T., Ishihara, D., et al., 2004, ApJ, 609, 203

Shimonishi, T., Onaka, T., Kato, D., et al., 2010, A&A, 514, A12

Smith, J.D., Braine, B.T., Dale, D.A., et al., 2007, ApJ, 656, 770

Song, I.-O., Kerr, T.H., McCombie, J., & Sarre, P.J., 2003, MNRAS, 346, L1

Tanaka, M., Matsumoto, T., Murakami, H., et al., 1996, PASJ, 48, L53


LABORATORY INFRARED SPECTROSCOPY OF PAHS

J. Oomens1, 2

Abstract. The hypothesis that polyaromatic molecules are the carriersof the infrared interstellar emission bands has spurred the laboratoryspectroscopy of this class of molecules. Here we will give an overviewof the infrared spectroscopic methods that have been applied over thepast two decades to investigate the IR spectra of PAHs, their ions andrelated species.

1 Introduction

One of the earlier comparisons of a spectrum of the unidentified infrared (UIR)bands – from the Orion Bar region – and an experimental laboratory spectrum ofa sample of polyaromatic species was reported by Allamandola et al. (1985), whoused a low-resolution Raman spectrum of a car exhaust sample as the referencePAH spectrum. In their seminal work, Leger & Puget (1984) used a calculatedemission spectrum of coronene as a reference spectrum to explain the UIR bands.These examples illustrate the very limited availability of laboratory spectroscopicdata on PAH species at the time the PAHs were first proposed as UIR band carri-ers. Although some pellet and solution-phase spectra of PAHs had been obtainedwith grating instruments (Colangeli et al. 1992), the PAH hypothesis inspiredspectroscopists around the World to record infrared PAH spectra under betterdefined and more astrophysically relevant conditions.

The interstellar emission spectra are believed to be carried by gas-phase PAHspecies in complete isolation. Except in highly irradiated environments, theirrotational temperature is believed to be below 100 K (Ysard & Verstraete 2010),while the vibrational temperature is estimated to be on the order of 1000 K afterthe absorption of an interstellar UV photon and subsequent internal conversionreleasing the UV photon energy into the vibrational manifold of the electronicground state. Finally, spectra recorded in emission rather than in absorption areastrophysically more relevant.

1 FOM Rijnhuizen, Edisonbaan 14, 3439MN Nieuwegein, The Netherlands2 van’t Hoff Institute for Molecular Sciences, University of Amsterdam, Amsterdam,The Netherlands



In this contribution, we will briefly review the experimental methods that havebeen applied over the past 25 years to obtain infrared spectra of isolated PAHs.

2 Neutral PAH infrared spectroscopy

To reduce the effects of the medium on the IR spectra recorded (Flickinger et al.1991; Moutou et al. 1996), an inert cryogenic Ar matrix was used instead of the saltpellets or solution (typically CCl4). This also enabled spectroscopists to obtainspectra at low temperature, which is more relevant to the conditions of denseclouds. But even the rare gas matrix can influence the appearance of the IRspectra (Joblin et al. 1994) and hence gas-phase absorption spectra were recorded,although this required heating the samples in an oven as a consequence of the lowvapor pressure of the PAHs. The use of molecular beam methods combined withlaser spectroscopy allowed spectroscopists to obtain low-temperature gas-phaseabsorption spectra. It was further realized that emission spectra, as observedfrom the ISM, may be different from absorption spectra as recorded with theuse of (commercial) spectrometers. Various spectroscopists thus set out to recordgas-phase IR emission spectra, which perhaps provide the most astrophysicallyrelevant neutral PAH spectra to date.

2.1 Matrix isolation spectroscopy

In matrix isolation, a host molecule (the PAH in this case) and an inert guest(often the rare gas atoms Ar or Ne) are simultaneously condensed onto an opticalwindow, which is kept at a cryogenic temperature in a cryostat. The techniquehas been widely applied to study reactive species, which are stabilized by the lowtemperature and inert environment. As a consequence of the low temperatureand the quenching of rotational degrees of freedom in the matrix, IR spectra ofmatrix-isolated PAHs feature sharp absorption bands with FWHM linewidths onthe order of 1 cm−1. An extensive database of infrared PAH spectra recorded incryogenic Ar matrices has been collected by the NASA Ames group1.

Recent applications include the study of nitrogen containing PAHs (Mattiodaet al. 2003), which have been suggested to provide a better match for the inter-stellar 6.2 μm emission band (Hudgins et al. 2005; Peeters et al. 2002), and ofcomplexes of PAHs and atomic Fe, which have been suggested to be stable ininterstellar clouds (Klotz et al. 1996; Wang et al. 2007). The method has alsobeen applied to record PAH spectra in the far-infrared range of the spectrum(Mattioda et al. 2009), where very molecule-specific vibrational modes – the so-called “drumhead” modes – are located.

1See www.astrochem.org/pahdb

J. Oomens: Laboratory Infrared Spectroscopy of PAHs 63

2.2 Gas-phase absorption spectroscopy

While the matrix isolation spectra provide high resolution spectra for a largenumber of PAH molecules, the influence of the matrix on the frequencies and,in particular, intensities remains unknown (Joblin et al. 1994). Gas-phase spectrawere considered to be more relevant from an astrophysical point of view and could,moreover, shed more light on possible matrix effects. Except for the smallestmembers of the family, however, the vapor pressure of PAHs is too low to recorddirect absorption spectra.

This issue was addressed by Joblin et al. (1994), who set up a heatpipe oveninside an FTIR spectrometer to record gas-phase IR absorption spectra of PAHsup to the size of ovalene (C32H14). This study indicated that while gas-phase andmatrix isolated band positions reproduce to within about 1%, band intensities canchange by up to a factor of five. While the high temperatures needed to vaporizethe molecules lead to substantially larger bandwidths and may appear to be as-trophysically unrealistic, interstellar emission occurs from excited PAH moleculeswhich are estimated to possess vibrational temperatures on the order of 1000 K(Joblin et al. 1995). Moreover, spectra recorded at different temperatures of theoven yield direct information on the anharmonic frequency shifts of vibrationalbands (Pirali et al. 2009), which are important in modelling IR emission spectra(Cook & Saykally 1998) and which are hard to estimate computationally for largesystems.

Gas-phase spectra recorded at low temperature are expected to give the bestvalues for the intrinsic molecular band positions, free from shifting due to anhar-monic interactions or matrix effects. To obtain such low-temperature spectra forlow-vapor pressure compounds, molecular beam techniques can be applied. Here,low rovibrational temperatures are achieved by expansion cooling of the moleculesin a supersonic beam of rare-gas atoms. Typical rotational temperatures are onthe order of 10 K; vibrational cooling can be less efficient. The low densities ofmolecules in these beams require the application of tunable lasers to obtain (in-frared) spectra of the jet-cooled PAHs. Piest et al. (2001) applied such methodsat the infrared free electron laser facility FELIX (Oepts et al. 1995) to obtain IRspectra of small PAH molecules, such as phenanthrene.

2.3 Infrared emission spectra

Interstellar UIR spectra are observed as emission spectra rather than as absorp-tion spectra and hence, laboratory emission spectra are desirable. One of the maindifficulties in these experiments is to distinguish the PAH IR fluorescence from theblackbody emission of the environment. Early experiments on UV excited PAHspecies therefore only recorded spectra in the 3 μm spectral range (Shan et al.1991; Williams & Leone 1995). Thermally heated gas-phase PAH samples wereused to increase the number density of emitting molecules so that the infraredPAH emission could be better distinguished from the background allowing variousgroups to record spectra at higher resolution (Pirali & Vervloet 2006) and further


into the infrared (Joblin et al. 1995). The evaporated PAHs, however, form asample in thermal equilibrium, while UV-excited PAHs combine low rotationalexcitation with high vibrational excitation, a situation that more closely resem-bles the assumed conditions in the ISM. Perhaps the most sophisticated emissionspectroscopy experiments have been reported by the group of Saykally, who used acryogenically cooled grating spectrometer to greatly suppress the blackbody back-ground radiation, in combination with an extremely sensitive IR detector (Cooket al. 1998, 1996). Thus, emission spectra of UV excited PAHs were recorded overthe entire IR range from 3 to 15 μm.

3 IR spectroscopy of ionic PAH species

One of the main conclusions of the emission spectra of Cook et al. (1996) wasthat neutral PAH species would not be able to explain the UIR emission featuresand the presence of a substantial fraction of ionized PAHs was invoked. In fact,already early on in the history of the PAH hypothesis it was realized that ionizedPAHs may play an important role because of the low ionization potentials ofPAHs combined with the UV pump fields that are necessary to induce IR emission(Allamandola et al. 1989). An important difference between the spectra of neutraland cationic PAHs was first revealed by quantum-chemical calculations on thenaphthalene radical cation by Pauzat et al. (1992); see also the contribution byPauzat elsewhere in this volume.

These considerations inspired various laboratory spectroscopists to record spec-tra of ionized PAH species, again preferably under astrophysically relevant con-ditions. Because of their mutual repulsion and inherently low gas-phase numberdensities, direct absorption spectra of (mass-selected) gas-phase ions are difficult, ifnot impossible, to record. Matrix isolation spectroscopy offers a way around thisproblem, but also gas-phase “action spectroscopy” schemes have more recentlybeen applied. Here we discuss the different methods applied to date and comparetheir merits and limitations.

3.1 Matrix isolation spectroscopy

Immediately following the computations of Pauzat et al. (1992), the group of Valaat the University of Florida set out to record the naphthalene cation spectrumusing a modified matrix isolation set up (Szczepanski et al. 1992). The effusivenaphthalene beam, which is co-deposited onto the 12-K BaF2 substrate with anAr beam, was intercepted by an electron beam with electron energies on the or-der of 50 – 200 eV. The Ar sample was premixed with CCl4, which acts as anelectron acceptor, reducing the re-neutralization of the naphthalene cations oncein the matrix. The IR spectrum of the matrix was then recorded using an FTIRspectrometer. One possible pitfall of this method is the possible recombinationor reactivity of the naphthalene cation in the matrix. Observed IR bands of neu-tral naphthalene are therefore subtracted. Moreover, to assign newly observed IRbands to the cation of interest, their intensity is typically correlated with those of


known bands in a UV/vis spectrum recorded from the same matrix (Szczepanski& Vala 1991).

The experimental naphthalene cation spectrum confirmed the computationalresults and showed definitively that the relative intensities of the different fami-lies of vibrational modes undergo major changes upon ionization. Soon after thisfirst PAH cation spectrum, IR matrix isolation spectra of a whole series of cationicPAHs were reported mainly by the Florida (Szczepanski & Vala 1993a) and NASAAmes (Hudgins & Allamandola 1995a,b; Hudgins et al. 1994) groups. These spec-tra provided the first experimental evidence that ionized PAHs may be abundantin the ISM (Allamandola et al. 1999; Szczepanski & Vala 1993b). In addition,these spectra provided the experimental verification for various quantum-chemicalstudies on the IR properties of ionized PAHs. For the vast majority of the PAHspecies that have been studied, a good correlation was found between experimen-tal frequencies (and intensities) and values computed at the density functionallevel of theory (see e.g. Langhoff (1996)). This provides confidence in the spectracalculated for species that are experimentally not easily accessible, such as verylarge PAHs (Bauschlicher et al. 2008; Malloci et al. 2007).

Despite the success of matrix isolation spectroscopy and the ever growingdatabase of spectra, the possible unknown effects of the matrix on band frequen-cies and intensities made that the desire to record true gas-phase PAH ion spectraremained. The main hurdle to record such spectra were the extremely low densitiesin which gas-phase ions are typically produced as a consequence of their mutualCoulombic repulsion. Spectroscopic methods based on conventional direct absorp-tion techniques are therefore generally not applicable to (mass-selected) molecularions. Since the 1980’s, however, infrared laser-based methods became increasinglysuccessful in recording gas-phase molecular ion spectra (Duncan 2000).

3.2 Messenger atom spectroscopy

One of the most influential laser-based methods was originally introduced by Y.T.Lee and coworkers and became known as “messenger-atom” spectroscopy (Lisy2006; Okumura et al. 1985, 1990). The method is based on a supersonic expansionin which ionized molecular species form weakly bound complexes with neutralmolecules or atoms, which are mass-selected typically in a quadrupole mass selector(QMS) or sector magnet. Upon excitation with an IR laser, the complex undergoespre-dissociation and the now bare molecular ion is detected in a mass analyzer,typically a time-of-flight (TOF) or quadrupole mass spectrometer. In all suchcombinations of tandem mass spectrometry and laser spectroscopy, monitoringthe intensity of the mass peak of the product molecular ion as a function of the IRwavelength then generates an IR spectrum of the ionic complex. While the entireionic complexes have been the subject of many studies, it was also realized that theneutral partner in the complex could be used solely as the detector of IR photonabsorption by the molecular ion, as signalled by its detachment from the ion, hencethe term “messenger”. Clearly in this case, the messenger was to have an as smallas possible influence on the ionic species of interest – or in the words of Y.T. Lee,


it was to behave as a spy (Okumura et al. 1990) – and hence, the use of rare gasatoms rather than small molecules became popular. Messenger atom spectroscopywas for instance used to record the IR spectra of several ionized aromatic systems,particularly benzene and various of its ring-substituted derivatives (Solca & Dopfer2004).

The early studies based on messenger spectroscopy focused particularly on thehydrogen stretching vibrations in the 3 μm wavelength range, since tunable ra-diation can be generated in this range by various methods, including frequencymixing of the output of a tunable dye laser in birefringent crystals. Towards thelate 1990’s, new widely tunable laser sources became available that covered theastrophysically interesting wavelength range roughly between 6 and 16 μm. Novelbirefringent materials allowed non-linear frequency down-conversion methods topenetrate further into the IR (Bosenberg & Guyer 1993). Moreover, tunable in-frared free electron lasers (O’Shea & Freund 2001), where radiation is generatedby a relativistic beam of electrons injected into a periodic magnetic field structure(wiggler), entered the realm of spectroscopy.

Using the free electron laser at the FELIX facility, Meijer and coworkersrecorded the gas-phase IR spectra of the naphthalene (Piest et al. 1999a) andphenanthrene (Piest et al. 2001) cations using a variation of the messenger-atomspectroscopy method (Piest et al. 1999b). Van der Waals clusters of the PAH anda rare gas atom (Ar in most cases) were produced in a pulsed molecular beamexpansion and they were resonantly ionized using a two-color resonance enhancedmultiphoton ionization (REMPI) scheme. The thus formed intact PAH+-Ar com-plexes were then irradiated by the output of the free electron laser, which wastuned in wavelength across the entire 6 to 16 μm range. IR absorption was sig-nalled as the appearance of the bare PAH+ ion in a TOF mass spectrometer.Since the influence of the weakly bound Ar atom is assumed to be negligible, thespectra, which feature a resolution of just a few cm−1, can be considered as beingdue to the cold and isolated PAH cation.

The group of Duncan recently recorded IR spectra of protonated benzene(Douberly et al. 2008) and naphthalene (Ricks et al. 2009) using an optical para-metric oscillator (OPO) source that provides mJ pulse energies down to wave-lengths of about 10 μm. Argon tagged protonated molecules are produced in theexpansion from a pulsed discharge nozzle and mass-selected in the first arm of areflectron TOF mass spectrometer. The ions are then exposed to the IR radiationin the turning point of the reflectron stack and mass analysis is accomplished inthe second arm of the reflectron. Also here, detachment of the messenger (Ar) isused to signal IR absorption by the protonated PAH, giving the spectrum of theionized PAH at a resolution of a few cm−1.

3.3 IR multiple photon dissociation spectroscopy

In the 1980’s, various groups explored the possibilities of gas-phase infrared ionspectroscopy by wavelength selective multiple photon dissociation of ions in storagemass spectrometers using powerful line-tunable CO2-lasers (see Eyler 2009 for a


recent review). Although a variety of ionic systems was successfully studied, thelimited tuning range of the CO2-laser (roughly 9 – 11 μm) severely impeded thepractical applicability of this method for spectroscopic purposes. More recently,the infrared pulse energies produced by free electron lasers were shown to behigh enough to induce photodissociation of vapor-phase PAH cations (Oomenset al. 2000). Installation of an ion trap time-of-flight mass spectrometer at thefree electron laser facility FELIX thus allowed Oomens et al. (2003) to recordIR multiple photon dissociation (IRMPD) spectra for a large number of massselected ionic PAHs. The PAHs are vaporized in an oven and effuse to the centerof a quadrupole ion trap, where they are non-resonantly ionized by a UV laser(excimer). Possible UV-induced fragments are mass-selectively removed from thetrap and the isolated parent ions are irradiated with the FEL beam. The IRinduced fragments and the remaining parent ions are then mass-analyzed in aTOF spectrometer, so that an IRMPD spectrum can be reconstructed by plottingthe fragment yield as a function of the FEL wavelength.

Dissociation thresholds for PAH ions are typically in the 6 – 8 eV range, whichrequires the absorption of tens to hundreds of IR photons. In polyatomic moleculesof the size of PAHs, infrared multiple photon excitation typically proceeds viarapid redistribution of the vibrational energy from the excited mode into the bathformed by all vibrations of the molecule. Intramolecular vibrational redistribu-tion (IVR) thus allows the molecule to sequentially absorb multiple photons onthe same vibrational transition, because the upper level is constantly being de-excited by quenching to the bath states (Black et al. 1977; Eyler 2009; Grant et al.1978). IVR-mediated multiple photon excitation thus avoids vibrational ladderclimbing, which is an inefficient process due to the anharmonicity bottleneck. Themultiple-photon nature of this spectroscopic technique is reflected in the shapeof the spectral features observed, most notably from the increased bandwidth ascompared to messenger-atom or matrix isolation spectroscopy. The PAH ion spec-tra recorded at FELIX using this method display FWHM bandwidths on the orderof 30 cm−1 (Oomens et al. 2000).

At the free electron laser facilities CLIO and FELIX, the FEL beam has nowbeen coupled to powerful tandem mass spectrometers, so that a variety of ionicPAH-related species can be spectroscopically studied. The ions can be producedby a variety of ion sources, including electrospray ionization (ESI) and laser des-orption/ablation methods. Much recent interest has been on protonated PAHspecies including protonated regular PAHs (Knorke et al. 2009; Lorenz et al. 2007;Vala et al. 2009b) and protonated N-containing PAHs (Alvaro Galue et al. 2010;Vala et al. 2009a). Clear differences are observed in the spectra of (even-electron)protonated species and (odd-electron) radical cation species. PAHs have longbeen suggested to efficiently bind to astrophysically abundant metal atoms or ions(Joalland et al. 2009; Klotz et al. 1996) and such species are now being investigatedspectroscopically using IRMPD spectroscopy (Simon et al. 2008; Szczepanski et al.2006).


3.4 IR emission spectroscopy

As mentioned in Section 2.3 for neutral PAH species, IR spectra recorded in emis-sion instead of absorption, as is more common, have the advantage that theymimick better the interstellar situation, where the UIR bands are observed inemission as well. However, the difficulty in recording emission spectra in theastrophysically most relevant 6 – 16 μm range lies mainly in distinguishing theweak infrared PAH fluorescence from the blackbody radiation of the environment.This is an even more daunting task for ionic species as compared to neutrals, asthe ion densities are intrinsically much lower. Nonetheless, the Saykally group atBerkeley managed to adapt their cryogenically cooled spectrometer so that it couldbe used to record emission spectra of gas-phase ionic PAHs (Kim & Saykally 2003).To this end, the spectrometer is coupled to an electron impact ion source, whichproduces high ion currents of more than 6 μA. Electron impact with approxi-mately 70 eV electrons not only ionizes the molecules, but also induces vibronicexcitations that are comparable to UV excitation; hence, the IR emission fromthe ions is assumed to be similar to that from UV excited ions as occurs underinterstellar conditions. Using a quadrupole deflector, the ions are injected intoa reflectron, which is placed in front of the entrance slit of the spectrometer. Areflectron is used to increase the residence time of the ions in the viewing regionof the spectrometer. All ion optics in the line of sight of the spectrometer need tobe cryogenically cooled while being electrically isolated, which constituted one ofthe main design challenges (Kim & Saykally 2003).

IR emission spectra were reported for the pyrene cation (C16H+10) and several

dehydrogenated derivatives (Kim & Saykally 2002; Kim et al. 2001). The pyrenecation spectrum showed a good correspondence with matrix isolation and IRMPDspectra of the pyrene cation. Additional bands observed in the emission spectrumcould be attributed to partly dehydrogenated fragments of the pyrene cation.

4 Experimental methods compared

A discussion of the various experimental methods applied for PAH spectroscopyinevitably raises the question as to which of the techniques is superior. Thisquestion is, however, not easily answered: each method has its strong and weakpoints.

Of all spectroscopic methods discussed here, IR emission spectroscopy yieldsspectra under conditions that are arguably most closely mimicking those of theinterstellar PAHs. The gaseous PAHs are electronically excited to energies inthe range of the interstellar conditions. The effects of vibrational anharmonicity(Barker et al. 1987; Pirali et al. 2009) are naturally incorporated into the ob-served emission bands. This is clearly an advantage over spectra obtained usingabsorption spectroscopy since for a true comparison to interstellar emission spec-tra, one would have to convolute the bands using an emission model including thevibrational anharmonicities (Cook & Saykally 1998; Pech et al. 2002), which arehowever largely unknown. However, the experiments are extremely challenging,


particularly for ionic PAHs. The method requires a cryogenically cooled IR spec-trometer with a very sensitive IR photon detector; for the ions, a high ion currentsource is required as well. Thus far, only the group of Saykally have reported IRemission spectra of PAH ions in the 5 – 15 μm spectral range, and only very fewspectra (of pyrene and some of its dehydrogenated fragments) have been reported.

Quite in contrast, matrix isolation spectroscopy is experimentally most readilyimplemented. Indeed, most PAH ion spectra reported to date have been obtainedusing this method. We may also note that variants of this method are currentlybeing applied to study the spectroscopy and (photo-)chemistry of PAHs and otherspecies in interstellar ices (Bernstein et al. 2007; Bouwman et al. 2009; Klotz et al.1996). Matrix isolation spectra feature narrow bandwidths induced by the lowtemperature and the quenching of rotational motion in the rare gas matrix. Ex-perimentally, matrix isolation spectroscopy may suffer from unknown effects ofthe matrix on the spectrum; in the infrared spectra, it appears that particularlyrelative intensities are affected (Joblin et al. 1994). Moreover, there is some uncer-tainty as to the exact nature of the species trapped in the matrix, particularly uponelectron or UV irradiation, and correlation measurements are usually required toassign bands to specific species (Szczepanski & Vala 1991). This problem may bepartly circumvented by mass-selective deposition of ions created ex-situ, see e.g.Fulara et al. (1993).

PAH ion spectra obtained using messenger atom spectroscopy have arguablyprovided the most reliable linear absorption spectra thus far. Bandwidths of just afew cm−1 have been obtained for PAH species that are estimated to be expansioncooled to temperatures on the order of 20 K. These methods require the availabilityof IR laser sources that are continuously tunable in the 5 – 15 μm spectral rangeand thus far free electron lasers and optical parametric oscillators have been used.Moreover, the PAHs need to be entrained in a seeded molecular beam expansionto form the weakly bound PAH-rare gas complexes, which has thus far only beenpossible for the smallest PAHs naphthalene (Piest et al. 1999a; Ricks et al. 2009)and phenanthrene (Piest et al. 2001).

IRMPD spectroscopy has the advantage that it can be conveniently coupledto (commercial) tandem mass spectrometers, providing access to many speciesbeyond regular PAH ions. A tunable laser source with high output powers isrequired to induce fragmentation of the ions and thus far, mainly the free elec-tron lasers FELIX in The Netherlands (Oomens et al. 2000) and CLIO in France(Lorenz et al. 2007) have been used for these studies. The spectral band shapes ob-served are clearly influenced by the process of multiple photon excitation (Oomenset al. 2003) resulting in FWHM bandwidths of typically 30 cm−1 or more, whichtends to produce unresolved spectra for larger, more irregularly shaped species.Nonetheless, PAH ion spectra up to the size of coronene (C24H12) have beenreported (Knorke et al. 2009; Oomens et al. 2001; Simon et al. 2008). In addi-tion, one may note that the broadening of bands is (in part) caused by vibra-tional anharmonicity; although the processes of IR multiple photon excitation andemission are different, the bandwidths observed in the spectra are comparable(Oomens et al. 2003). The use of different ion sources and the capability of mass


selection prior to spectroscopic interrogation makes this method very versatile,which is reflected in the variety of ionic PAH-related species that have been inves-tigated, such as Fe+PAH complexes (Simon et al. 2008; Szczepanski et al. 2006)and protonated PAHs (Knorke et al. 2009; Lorenz et al. 2007; Vala et al. 2009b).We anticipate that mass spectrometry based spectroscopic methods will be ideallysuited to investigate structures of PAH photo-products (Ekern et al. 1997; Jochimset al. 1994), which may provide important information to better understand thechemistry of PAHs in the diffuse ISM.

New technologies to record IR spectra of ionized species are continuously beingexplored and they may hold great promise for improved PAH ion spectra. Usingvarious spectroscopic schemes, cryogenically cooled multipole ion traps have beenrecently applied to record spectra of cold gas-phase ions. Laser-induced reactionshave been applied to obtain IR spectra of small hydrocarbon ions of astrophys-ical interest, including the intriguing CH+

5 (Asvany et al. 2005) and hydronium(Asvany et al. 2008) ions. Cold ion traps have been applied to record electronicspectra of carbon chain ions of astrophysical interest (Rice et al. 2010). Combi-nation of UV and IR lasers in cold ion traps to obtain IR hole-burning spectraof the cryogenically cooled ions has been applied to biomolecules of various sizes(Nagornova et al. 2010; Stearns et al. 2007). We anticipate that such methods willbe applicable to PAH species as well.

Superfluid nanoscopic He droplets are known to provide an extremely non-interacting and cold (0.4 K) matrix to study the spectroscopy of molecular species(Goyal et al. 1992; Hartmann et al. 1996). Drabbels and coworkers are now apply-ing these methods to study the spectroscopy of ions, including aromatic species,at high resolutions (Loginov et al. 2008; Smolarek et al. 2010).

5 Concluding remarks

In this review, I have given an overview of the various experimental methods thathave been applied to obtain IR spectra of PAHs since the 1980’s, when thesespecies were originally hypothesized to occur abundantly in the ISM. The exper-imental IR spectra provide important benchmarks for theoretical investigationsof the spectral properties of PAHs, which are discussed elsewhere in this volume.In addition, theory and experiments to understand PAH spectra in other spectralranges, particularly in the UV/vis range are addressed elsewhere in this issue. Wethus conclude that much progress has been made over the past 25 years in theexperimental characterization of the spectra of PAHs, their ions and various oftheir derivatives.

References

Allamandola, L., Hudgins, D., & Sandford, S., 1999, Astrophys. J., 511, L115

Allamandola, L., Tielens, A., & Barker, J., 1985, Astrophys. J., 290, L25

Allamandola, L., Tielens, A., & Barker, J., 1989, Astrophys. J. Suppl. Ser., 71, 733


Alvaro Galue, H., Pirali, O., & Oomens, J., 2010, Astron. Astrophys., 517, A15

Asvany, O., Ricken, O., Muller, H.S.P., et al., 2008, Phys. Rev. Lett., 100, 233004

Asvany, O., Schlemmer, S., Redlich, B., et al., 2005, Science 309, 1219

Barker, J., Allamandola, L., & Tielens, A., 1987, Astrophys. J., 315, L61

Bauschlicher, C.W., Peeters, E., & Allamandola, L.J., 2008, Astrophys. J., 678, 316

Bernstein, M.P., Sandford, S.A., Mattioda, A.L., & Allamandola, L.J., 2007, Astrophys.J., 664, 1264

Black, J.G., Yablonovitch, E., Bloembergen, N., & Mukamel, S., 1977, Phys. Rev. Lett.,38, 1131

Bosenberg, W.R., & Guyer, D.R., 1993, J. Opt. Soc. Am. B, 10, 17161722

Bouwman, J., Paardekooper, D.M., Cuppen, H.M., Linnartz, H., & Allamandola, L.J.,2009, Astrophys. J., 700, 56

Colangeli, L., Mennella, V., Baratta, G.A., Bussoletti, E., & Strazzulla, G., 1992, Astro-phys. J., 396, 369

Cook, D., & Saykally, R., 1998, Astrophys. J., 493, 793

Cook, D., Schlemmer, S., Balucani, N., et al., 1998, J. Phys. Chem. A, 102, 1465

Cook, D., Schlemmer, S., Balucani, N., et al., 1996, Nature, 380, 227

Douberly, G.E., Ricks, A.M., Schleyer, P.v.R., & Duncan, M.A., 2008, J. Phys. Chem.A, 112, 4869

Duncan, M.A., 2000, Int. J. Mass Spectrom., 200, 545

Ekern, S., Marshall, A., Szczepanski, J., & Vala, M., 1997, Astrophys. J., 488, L39

Eyler, J.R., 2009, Mass Spectrom. Rev., 28, 448

Flickinger, G.C., Wdowiak, T.J., & Gomez, P.L., 1991, Astrophys. J., 380, L43

Fulara, J., Jakobi, M., & Maier, J., 1993, Chem. Phys. Lett., 211, 227

Goyal, S., Schutt, D.L., & Scoles, G., 1992, Phys. Rev. Lett., 69, 933

Grant, E.R., Schulz, P.A., Sudbo, A.S., Shen, Y.R., & Lee, Y.T., 1978, Phys. Rev. Lett.,40, 115

Hartmann, M., Miller, R.E., Toennies, J.P., & Vilesov, A.F., 1996, Science, 272, 1631

Hudgins, D., & Allamandola, L., 1995a, J. Phys. Chem., 99, 3033

Hudgins, D., & Allamandola, L., 1995b, J. Phys. Chem., 99, 8978

Hudgins, D., Bauschlicher, C., & Allamandola, L., 2005, Astrophys. J., 632, 316

Hudgins, D., Sandford, S., & Allamandola, L., 1994, J. Phys. Chem., 98, 4243

Joalland, B., Simon, A., Marsden, C.J., & Joblin, C., 2009, Astron. & Astrophys., 494,969

Joblin, C., Boissel, P., Leger, A., d’Hendecourt, L., & Defourneau, D., 1995, Astron.Astrophys., 299, 835

Joblin, C., d’Hendecourt, L., Leger, A., & Defourneau, D., 1994, Astron. Astrophys.,281, 923

Jochims, H.W., Ruhl, E., Baumgartel, H., Tobita, S., & Leach, S., 1994, Astrophys. J.,420, 307

Kim, H., & Saykally, R., 2002, Astrophys. J. Suppl. Ser., 143, 455

Kim, H., & Saykally, R., 2003, Rev. Sci. Instrum., 74, 2488

Kim, H., Wagner, D., & Saykally, R., 2001, Phys. Rev. Lett., 86, 5691


Klotz, A., Marty, P., Boissel, P., et al., 1996, Plan. Space Sci., 44, 957

Knorke, H., Langer, J., Oomens, J., & Dopfer, O., 2009, Astrophys. J., 706, L66

Langhoff, S.R., 1996, J. Phys. Chem., 100, 2819

Leger, A., & Puget, J., 1984, Astron. Astrophys., 137, L5

Lisy, J.M., 2006, J. Chem. Phys., 125, 132302

Loginov, E., Braun, A., & Drabbels, M., 2008, Phys. Chem. Chem. Phys., 10, 6107

Lorenz, U.J., Solca, N., Lemaire, J., Maitre, P., & Dopfer, O., 2007, Angew. Chem. Int.Ed., 46, 6714

Malloci, G., Joblin, C., & Mulas, G., 2007, Chem. Phys., 332, 353

Mattioda, A., Hudgins, D., Bauschlicher, C., Rosi, M., & Allamandola, L., 2003, J. Phys.Chem. A, 107, 1486

Mattioda, A.L., Ricca, A., Tucker, J., Bauschlicher, Jr., C.W., & Allamandola, L.J.,2009, Astron. J., 137, 4054

Moutou, C., Leger, A., & dHendecourt, L., 1996, Astron. Astrophys., 310, 297

Nagornova, N.S., Rizzo, T.R., & Boyarkin, O.V., 2010, J. Am. Chem. Soc., 132, 4040

Oepts, D., van der Meer, A.F.G., & van Amersfoort, P.W., 1995, Infrared Phys. Technol.,36, 297

Okumura, M., Yeh, L.I., & Lee, Y.T., 1985, J. Chem. Phys., 83, 3705

Okumura, M., Yeh, L.I., Myers, J.D., & Lee, Y.T., 1990, J. Phys. Chem., 94, 3416

Oomens, J., Sartakov, B.G., Tielens, A.G.G.M., Meijer, G., & von Helden, G., 2001,Astrophys. J., 560, L99

Oomens, J., Tielens, A.G.G.M., Sartakov, B.G., von Helden, G., & Meijer, G., 2003,Astrophys. J., 591, 968

Oomens, J., van Roij, A.J.A., Meijer, G., & von Helden, G., 2000, Astrophys. J., 542,404

O’Shea, P.G., & Freund, H.P. 2001, Science, 292, 1853

Pauzat, F., Talbi, D., Miller, M.D., DeFrees, D.J., & Ellinger, Y., 1992, J. Phys. Chem,96, 7882

Pech, C., Joblin, C., Boissel, P., 2002, Astron. Astrophys., 388, 639

Peeters, E., Hony, S., Van Kerckhoven, C., et al., 2002, Astron. Astrophys., 390, 1089

Piest, H., von Helden, G., & Meijer, G., 1999a, Astrophys. J., 520, L75

Piest, H., von Helden, G., & Meijer, G., 1999b, J. Chem. Phys., 110, 2010

Piest, J., Oomens, J., Bakker, J., von Helden, G., & Meijer, G., 2001, Spectrochim. ActaA, 57, 717

Pirali, O., & Vervloet, M., 2006, Chem. Phys. Lett., 423, 376

Pirali, O., Vervloet, M., Mulas, G., Malloci, G., & Joblin, C., 2009, Phys. Chem. Chem.Phys., 11, 3443

Rice, C.A., Rudnev, V., Dietsche, R., & Maier, J.P., 2010, Astron. J., 140, 203

Ricks, A.M., Douberly, G.E., & Duncan, M.A., 2009, Astrophys. J., 702, 301

Shan, J., Suton, M., & Lee, L.C., 1991, Astrophys. J., 383, 459

Simon, A., Joblin, C., Polfer, N., & Oomens, J., 2008, J. Phys. Chem. A, 112, 8551

Smolarek, S., Vdovin, A., Perrier, D.L., et al., 2010, J. Am. Chem. Soc., 132, 6315

Solca, N., & Dopfer, O., 2004, J. Am. Chem. Soc., 126, 1716


Stearns, J.A., Mercier, S., Seaiby, C., et al., 2007, J. Am. Chem. Soc., 129, 11814

Szczepanski, J., Roser, D., Personette, W., et al., 1992, J. Phys. Chem., 96, 7876

Szczepanski, J., & Vala, M., 1991, J. Phys. Chem., 95, 2792

Szczepanski, J., & Vala, M., 1993a, Astrophys. J., 414, 646

Szczepanski, J., & Vala, M., 1993b, Nature, 363, 699

Szczepanski, J., Wang, H., Vala, M., et al., 2006, Astrophys. J., 646, 666

Vala, M., Szczepanski, J., Dunbar, R., Oomens, J., & Steill, J.D., 2009a, Chem. Phys.Lett., 473, 43

Vala, M., Szczepanski, J., Oomens, J., & Steill, J.D., 2009b, J. Am. Chem. Soc., 131,5784

Wang, Y., Szczepanski, J., & Vala, M., 2007, Chem. Phys., 342, 107

Williams, R.M., & Leone, S.R. 1995, Astrophys. J., 443, 675

Ysard, N., & Verstraete, L. 2010, Astron. & Astrophys., 509, A12


COMPUTATIONAL IR SPECTROSCOPY FOR PAHS: FROMTHE EARLY YEARS TO THE PRESENT STATUS

F. Pauzat1

Abstract. In the long story of interstellar PAHs, computations haveplayed and are still playing a fundamental role in connection with ex-periments and observations. From the very first calculations of the IRspectra of small PAHs in the late eighties to the more recent ones, ev-ery aspect of the research linked to the PAH hypothesis has evolveddramatically: the size and the variety of the molecules considered, thetechniques used, the precision of the astronomical observations . . . Theinitial landscape has completely changed though the quest is still thesame, that is to correlate the so-called UIR bands spectra ubiquitousin the ISM (Inter Stellar Medium) with a chemical family of molecules,the PAHs. An historical review of the 25 years of this quest is presentedhere, focusing on the computational part.

1 Introduction

Theoretical studies of chemical species and processes in extreme physical condi-tions are a clear need of the astrophysical community confronted with situationsalmost impossible to reproduce in laboratory experiments. Many questions arisethat do not find experimental answers easily. However, some of them can be ob-tained from theoretical studies that are the alternative in a number of problems.

One of these questions, which has proved to be a formidable challenge in mod-ern astrophysics, is the interpretation of the infrared emission bands of our galaxybetween 2 and 15 μm. These bands, referred to as “Unidentified Infra Red (UIR)bands” in the literature, were first discovered by ground-based observations at theend of the seventies. Since then, their existence has been confirmed in a numberof objects by observations using the IRAS, ISO and more recently Spitzer andAKARI space missions. This emission, characterized by a series of bands that aresystematically found at 3.3, 6.2, 7.7, 8.6, 11.3 and 12.7 μm, seems to be widelypresent in various environments of the interstellar medium (ISM). For interpreting

1 UPMC Univ. Paris 06, UMR - CNRS 7616, LCT, 75005 Paris, France



this emission, Leger & Puget (1984), Allamandola et al. (1985) proposed a modelbased on the similarity between the spectra observed in space and those of neutralpolycyclic aromatic hydrocarbons (PAH) obtained in the laboratory. In space,these molecules could become vibrationally excited by absorption of UV photonsand then release their excess energy by means of infrared fluorescence (Duley &Williams 1981). This is generally referred to as the PAH hypothesis.

Contrary to radio spectra whose interpretation relies on a perfect match be-tween observations and measurements on a single molecule, the observed IR spec-tra correspond to families of molecules so that the details of the spectral matchcannot be satisfactory; furthermore, the relative intensities of the different bandsobserved may vary significantly from an astronomical object to another. Vari-ous interpretations of these differences have been proposed, leading sometimes tocontradictory conclusions such as severe dehydrogenation or negligible dehydro-genation of the interstellar PAHs. It has been said that PAHs could be eitherneutral or positively charged due to ionization by the radiation field and that theycould be present in the form of negative ions within dense clouds (Wakelam &Herbst 2008). Given such a diversity of hypotheses, the ideal strategy would havebeen to measure the IR emission of a selected number of representative species inthe gas phase, at temperatures and densities reproducing the interstellar condi-tions in the laboratory. At that time, it was considered as a practically impossibletask, and numerical simulation was the alternative. The main difficulties to beaddressed were:

i) The size of molecules like PAHs, ranging from a few tenths to several hun-dreds of atoms, which was, when the study began, non tractable for mostab initio quantum chemistry codes.

ii) The number of different types of molecules (ramified, substituted, ionized,hydrogenated, de-hydrogenated, metallic-complexes . . .) to be treated at thesame level of accuracy, implying a huge amount of computer time.

iii) The level of accuracy necessary to connect calculations with observations,which raises two more problems. The first one refers to a real theoreticalchallenge: one has to compare an emission spectrum (observed) to an absorp-tion spectrum (computed). The second one is not relevant of pure quantumchemistry: the reference being the observed spectrum, i.e. a global spec-trum emitted by a mixture of molecules from a known family but with anunknown composition, the comparison depends of the composition chosenand there is no theoretical criterion to help designing a unique solution.

The difficult quest for theoretical IR spectra usable by astrophysicists, has beenextending from the last eighties to nowadays with successes as well as setbacks.The presentation will be split into three chronological parts. The first part retracesa “learning period”, beginning in the late eighties, with methods whose reliabilityto efficiency ratio was still to be optimized for such molecules and variables. Theoptions considered at the time are presented and discussed. It allowed to set up

F. Pauzat: Computational IR Spectroscopy for PAHs 77

standard procedures and eventually led to some spectacular results. The secondpart deals with the following period which, for the computation side, turned intoa systematic study of all possible PAHs. This mass production was undertakento get databases available to the PAH community, in order to combine the effortsof theoreticians and experimentalists (cf. Sect. 3.1). Indeed, mixing the results ofcalculations and of experiments turned out to be very fruitful. On top of thosecalculations, a large amount of work has been performed to produce syntheticspectra with variations of the type and size of the PAHs in order to comparedirectly with the observations and to deduce information about the populationas a function of the environment (see for instance Cami et al. in this volume).The third part reports the main theoretical advances made during the last years,aiming at a much more precise insight into the spectroscopy of PAHs. The ongoingdevelopments of new methods and the comparison of their results to astronomicalspectra with fine structures of increasing complexity can be regarded as a realchallenge. Going beyond the harmonic approximation and taking into accountthe distinction between emission and absorption spectra are the main points nowaddressed (see for example Basire et al. in this volume).

2 First generation computations: HF vs. DFT electronic calculations

The spectra observed between 2 and 50 μm are the signatures of molecular vi-brations. They are characterised by two quantities, namely, their frequencies andtheir intensities. The frequencies are the energetic signatures of the deformationsof the molecular structure whereas the intensities show how the electronic densityis distorted with the internal displacements of the nuclei. Together, they consti-tute a real fingerprint of the molecule, often used for identification in terrestriallaboratories.

Any non-linear molecule composed of N atoms can move along 3N degrees offreedom, that can be separated into 3 translations and 3 rotations which do notimply any modification of the geometry, and 3N-6 vibrations (reduced to 3N-5 inthe case of linear species) that involve periodic structural modifications with welldefined frequencies.

Due to the quantum nature of the particles involved, quantum chemistry ismost appropriate to describe the behavior of electrons and nuclei within a molecule.It is done by using mathematical functions capable of describing quantum phe-nomena from first principles and the universal constants of physics. Whatever thechoices made, evaluation of wave functions or evaluation of electronic densities, allmethods are based on solving the Schrodinger equation in which all the particlesand all the interactions between them are taken into account. It is expressed asH(r, R)Ψ(r, R) = EΨ(r, R)where H(r, R) is the complete Hamiltonian of the system depending on the elec-tronic r and nuclear R coordinates. This operator is expressed as H(r, R) =Te + TA + VeA + Vee + VAA

where Te and TA are the kinetic energy operators for the electrons and nuclei,VeA and Vee the operators describing the Coulomb attraction between electrons


and nuclei and the repulsion between the electrons, VAA the repulsion between thenuclei.

Solving the Schrodinger equation is not a simple process since the energy tobe calculated has contributions from various types of movements, motions of elec-trons, vibrations and rotations of the nuclear frame, all entangled in a subtle way.Rigorously, the Schrodinger equation can be solved exactly only for a two-electronatom. In all other cases, it is necessary to make approximations. Levering on thefact that nuclei are thousands times heavier than electrons, the first and most nat-ural approximation is to uncouple the two types of motions (Born-Oppenheimerprinciple). Then, the calculations are carried out in the three following steps.

i) The first step is to determine the geometry of the molecule by minimizingits electronic energy as a function of geometrical parameters such as bonddistances, bond angles, dihedral angles. It is a pure electronic calculationin which the wave function in the Schrodinger electronic equation, Ψ(r, R),is a function depending explicitly of all the electronic variables (space andspin) and parametrically of the nuclei coordinates. Ψ(r, R) is developed asa linear combination of suitable functions, known as basis sets. Differenttypes of mathematical functions are currently used, as for example, atomicfunctions, plane waves or gaussian functions.

ii) The second step is to describe the movement of the nuclei in the potentialpreviously calculated at fixed positions of the nuclei. Assuming that eachvibration can be approximated by a harmonic oscillator, the so-called har-monic frequencies are obtained from the diagonalization of an effective forceconstant matrix built from the atomic masses and the second derivatives ofthe energy as a function of the displacements of the nuclei in the vicinityof their equilibrium positions. The eigenvalues lead to the harmonic vibra-tional frequencies whereas the eigenvectors provide the displacement vectorsof the nuclei (normal coordinates). Going beyond the harmonic descriptionwould imply the mapping of a multidimensional energy potential allowingthe calculation of higher order derivatives and inter-vibration coupling terms(see Sect. 4).

iii) The third step is the calculation of the vibrations intensities that are directlyrelated to the first multipole moment that has a non vanishing derivative withrespect to the normal coordinates previously obtained (Amos 1984; Helgakeret al. 1986).

2.1 Independent electron model versus correlated electron model

Back to the eighties, the methods of computational chemistry were already elab-orated and could be very efficient but only for small sized molecules (3–4 atoms).Confronted to PAHs that are much larger molecules (several tens of atoms) andto the demand for an accurate determination of their IR frequencies and intensi-ties that are wave function derivatives, quantum chemists could only use part of


their codes, taking into account the computer possibilities of the time. Therefore,the first calculations were performed at the Hartree-Fock (HF) level; then camethe Density Functional Theory (DFT) option, which is still the major theoreticalscheme used today for the treatment of PAHs up to over a hundred of atoms.

In HF theory, electrons are assumed to be independent particles, each onemoving in the average field of the others; there is no correlation between the elec-trons. Bringing correlation effects into the electronic wavefunction, i.e. takinginto account that the movement of one electron is dependent on the instantaneousmovement of the others, can be done in several ways. In post HF theories, correla-tion effects are introduced by mixing the ground state wavefunction with excitedconfigurations. The simplest approach is known as MP2 (Moller Plesset 2nd or-der) in which only doubly excited configurations are considered in a perturbationtreatment. In DFT, electrons are explicitely correlated through what is known asa correlation functional, which, by definition, makes it possible to limit the cal-culation to the ground state only. However, the exact correlation functional notbeing known, the DFT methods have to rely on approximate functionals. How theHF and DFT approaches are related is presented below.In the HF theory, the energy as function of the density matrix P, has the form1:EHF = V + < hP > +1/2 < PJ(P ) > −1/2 < PK(P ) >with:V : nuclear repulsion energy,< hP >: one-electron (kinetic + potential) energy1/2< PJ(P ) >: coulomb repulsion of the electrons–1/2< PK(P ) >: exchange energy.In the Kohn-Sham (KS) formulation of density functional theory, the energy hasthe form:EKS = V + < hP > +1/2 < PJ(P ) > +EX [P ] + EC [P ]with:EX [P ]: exchange functionalEC [P ]: correlation functional.Within the Kohn-Sham formulation, Hartree-Fock theory can be regarded as aspecial case of density functional theory, with:EX [P ] = −1/2 < PK(P ) > and EC = 0.Although the inclusion of electron correlation in DFT can be seen as a qualitativeleap in the description of electronic structures, one delicate issue remains whenusing DFT, namely, the choice of the exchange and correlation functionals. Somedepend only of the local value of the density (LDA functionals), others dependnot only of the local value of the density, but also of the derivative of this density(GGA functionals). Others, trying to take the best of the HF and DFT approachesuse a combination of both (Hybrid functionals). This explains that the results canbe rather different depending of the functional used. The most popular and widely

1The notation <>, known as the bra-ket Dirac formalism means that the quantity < Q > isobtained by integration over the wave function space.


employed functional for the calculation of IR spectra is known under the acronymof B3LYP (Becke 1993; Lee et al. 1988). The form initially proposed by Becke forthe exchange correlation functional depends on 3 parameters that are adjusted tofit molecular properties of a representative corpus, mostly of organic molecules,which explains the success of this approach in the PAH studies.

At that point, general textbooks (for example: Szabo & Ostlund 1982; Lee& Scuseria 1995; Koch & Holthausen 2001) may help the interested reader toget a better insight into the original developments forming the basis of quantumchemistry as it stands today.

2.2 Correcting the missing effects: Scaling procedures

As mentioned before, all electronic methods are approximate when dealing withcorrelation, and at least part of the correlation effect is missing. Most of the time,evaluation of the IR frequencies is performed within the harmonic approximation.Thus, two independent effects, correlation and anharmonicity, have to be corrected.In such situations, the most efficient way to proceed is to adjust the raw values asa whole by calibrating on well known experimental data. Several procedures havebeen employed:

- a direct scaling of the frequencies using a unique scale factor for the entirespectrum or a specific scale factor by frequency range or frequency type,

- an indirect scaling through the force constant matrix, scaling only the diag-onal terms (uniformly or not) or scaling also the non-diagonal terms.

The scaling factors depend essentially on the method and the basis set employed,and, in a lesser way, of the type of molecules considered (families). From theoriginal work by Scott & Radom (1996), a number of numerical experiments havebeen discussed in the literature (see: Yoshida et al. 2000; Merrick et al. 2007;Johnson et al. 2010; Bauschlicher & Langhoff 1997, for the specific case of PAHs).One has to note that, contrary to frequencies, no scaling procedure is available forintensities which are linked to derivatives of observables.

2.3 Successful applications

The approximations detailed above should not mask the fact that an impressiveamount of results has been obtained with these methods, allowing a real insightinto the PAH spectroscopic features.

For example, a major question was how to justify the weakness of the 3.3 μmfeature observed, compared to the corresponding experimental value (known onlyfor standard neutral PAHs at that time). Taking into account that this featurecorresponds to the CH stretching vibration, the first idea that came in was to arguein favour of dehydrogenated PAHs, assuming the CH intensity to be proportionalto the number of CH bonds. However, no spectrum of any dehydrogenated PAHwas available and no data could support such an assumption, though it seemed areasonable one.


Naphthalene spectra Cumulative spectra

Fig. 1. Absorption spectra of neutral versus ionized PAHs. Left: effects of dehydro-

genation and ionization in the naphthalene family (HF/321G calculations with scaling

factor specific of the type of vibration); the bands FWHM are the averaged widths of

the interstellar bands in the corresponding region of the spectra (adapted from Pauzat

et al. 1995). Right: histogram representations of an equally weighted sum of the IR

adsorption band intensities of 13 PAHs (B3LYP/4-31G with uniform scaling): neutrals

(top) and cations (bottom) (adapted from Langhoff 1996).


The first breakthrough came from quantum theory simulations showing thespectacular difference between neutrals and cations, in particular for the intensitiesof the 3.3 μm feature. Suggested in 1988 by HF calculations for the smallestPAHs (Communication by Miller & D. DeFrees, IAU Symp. 135, InterstellarDust, Santa Clara), it was confirmed in 1992 jointly by computations (Pauzatet al. 1992) and experiments (Szczepanski et al. 1992). Later on, more extensivestudies through HF (DeFrees et al. 1993) and DFT calculations (Langhoff 1996)as well as laboratory experiments emphasized this fact (see for example Oomensin this volume). The theoretical spectra reported in Figure 1 summarize the twoaspects of that important result. First the compilation of the spectra of a series of13 PAHs in their neutral and ionized forms (Langhoff 1996) illustrate the generalcollapse of the 3.3 μm feature obtained by ionisation. Second, the case study ofnaphthalene (Pauzat et al. 1995) demonstrates that the collapse is effectively dueto ionization of the PAH and not to some sort of dehydrogenation which wouldimply the same collapse of the 11.3 μm feature and consequently could not accountfor the observed intensity ratios. This fact has been confirmed for larger species(Pauzat et al. 1997).

It is worth mentioning that the first generation computations have been preciseand reliable enough to characterize the fluctuations in frequency positions andrelative intensities connected to almost all the different families of PAHs suggestedby the various interpretative models. Therefore, all sorts of PAH families havebeen calculated to support the interpretation of the observed IR spectra (see alsoPeeters in this volume). In addition to the large body of data obtained on neutraland singly ionized PAHs, more specific calculations were carried out on specificallytargeted families. A non exhaustive list of examples is given below focusing on thephysical effects, independently of the size of the PAH carriers. Among the effectsconsidered one may cite:

- multiple ionizations (Ellinger et al. 1999; Bauschlicher & Bakes 2000; Bakeset al. 2001; Malloci et al. 2007)

- negative ions (Langhoff 1996; Bakes et al. 2001; Bauschlicher et al. 2008,2009)

- dehydrogenated PAHs (Pauzat et al. 1995, 1997)- hydrogenated PAHs (Pauzat & Ellinger 2001; Bauschlicher et al. 2001; Bee-

gle et al. 2001; Ricks et al. 2009)- substitutions by lateral chains and carbon replacements by heteroatoms

(Langhoff et al. 1998; Pauzat et al. 1999; Hudgins et al. 2005)- irregular structures with 4/5/7 membered rings (Bauschlicher et al. 1999;

Pauzat & Ellinger 2002)- deuterated PAHs (Bauschlicher et al. 1997; Mulas et al. 2003; Peeters et al.

2004)- metal complexation (Fe, Mg, Si) (Ellinger et al. 1999; Cassam 2002; Hudgins

et al. 2005; Bauschlicher & Ricca 2009; Joalland et al. 2010; Simon & Joblin2007, 2010) . . ..


Stable calculation Unstable calculation

Closed shell neutral Open shell positive ion

Fig. 2. Examples of successful (left) and unstable (right) calculations of IR spectra

using BP86, B3LYP with the 4–31G basis set and FWHM of 30 cm−1 (adapted from

Bauschlicher et al. 1999).

2.4 Limitations

In spite of a number of successes and a satisfactory agreement with the experimen-tal spectra when available (apart from some lack of precision for the intensities),some theoretical problems remain to be solved with electronic calculations. In-stabilities, mostly due to symmetry breaking artefacts, may occur i.e. when thecalculated electronic density has a lower symmetry than the nuclear frame. Thishappens when two states fall quite close in energy in an open shell structure. Prac-tically, it generates abnormally large intensities as well as inaccurate positions ofthe bands, especially for in-plane asymmetric CC vibrations. Contrary to whatis sometimes advanced, this is true with all methods used for PAHs but hope-fully for a few molecules only. It is true for HF methods, especially when usingROHF for open-shells molecules (case of most cations), it is true for Moller Plesset


Perturbation theory (MP2) which has been used sometimes as reference for cor-relation evaluation, and it is also true of DFT though less often. In this last case,it seems to depend on the choice of the functional. One of those cases, the flu-oranthene cation, is discussed by Bauschlicher et al. (1999) considering B3LYPand BP86 DFT approaches (Fig. 2). However, the theoretically correct solutionof symmetry breaking in open shells relying on high level multiconfigurational cal-culations (McLean et al. 1985) is still out of reach today for most PAHs becauseof the size of the computational problem.

3 Second generation computations: Databases of PAHs IR spectra

3.1 PAHs databases

Thanks to a quickly increasing power of the computers, systematic calculations ofwhole families of PAHs, dealing with larger and larger PAHs (Bauschlicher 2002;Pathak & Rastogi 2007; Bauschlicher et al. 2008, 2009; Ricca et al. 2010), could becurrently performed. The time had come for a PAH business, meaning large scalecomputations implying significant human and computer-time potentials. Suchcharacteristics were available at NASA Ames Research Center. Moreover, exper-imental setups were available on the same site and cooperation between the twoproved very successful. The same objective was also present in the European col-laboration between the Cagliari Observatory and the CESR at Toulouse. Twodatabases are now currently on line:

The Cagliari database: astrochemistry.ca.astro.it/database/For more details, refer to the review by Malloci et al. (2007):

The NASA database: www.astrochem.org/pahdb/For more details, refer to the review by Bauschlicher et al. (2010) and see alsoBoersma et al. in this volume.

The exploitation of these on line databases of spectra is presented below.

3.2 Modeling interstellar composite IR spectra

Quantum Chemistry calculations give spectra of individual PAHs (free flyers) un-der a simple form: two numbers per vibration, the frequency and the correspondingabsolute intensity; they provide also a vector array which is the normal coordinatedescribing the movement associated to the vibration in the reference frame of themolecular geometry. It is the analysis of the normal coordinates in terms of elemen-tary displacements (stretching, bending...) that allows structural identification ofthe vibration. Starting from these data, one has to build a synthetic spectrum tobe compared to an observed spectrum. Two issues need to be addressed.

One has to determine the profile of each band to produce a full spectrumfor every PAH considered. Then, one has to figure which families of PAHs areconcerned, depending of the region of observation, and in which proportions theyare to be averaged.


Observation/Absorption/Emission Emission as a function of size

a

b

Fig. 3. Composite spectra: left – low-resolution spectrum of the reflection nebula

NGC 7023 (panel a) compared to the average absorption spectrum of 7 neutral and

singly positive ionised very large PAHs (panel b) and the computed emission spectrum

of the same PAH mixture at 465 K (panel c) (adapted from Boersma 2009); right – com-

parative emission spectra of PAHs having less than 30 C atoms (black) and more than

30 C atoms (grey) (a) neutrals and (b) cations (adapted from Pathak & Rastogi 2007).

In the case of a single PAH, it is now a priori possible to determine a theo-retical profile for each band through elaborate methods implying intramolecularvibrational redistribution, anharmonic shifts and hot bands, rotational broaden-ing, ... It is a highly sophisticated theoretical work, very demanding in man powerand computer time. This is hardly an option to be considered when dealing witha mass production of spectra for such oversized molecules as PAHs. Empiricalsubstitutes can be used instead.

Concerning the profile of a band, essentially two simplified representations areavailable:

- Histograms.

- Curves with Lorentzian / Gaussian shapes.

Then several options are possible for the FWHM (full width at half maximum):

- The frequency interval obtained for the calculated vibrations of the sametype


- A standard line width

- Different FWHM depending on vibration type and deduced fromexperiments.

The final step is relevant of pure astrochemistry only. The appropriate calcu-lated spectra have to be selected and averaged according variable percentages.Comparison of such composite spectra with the observed spectra allows to de-duce information about size, ionization or composition of the PAHs according theenvironment. Some examples of composite spectra are given in Figure 3.

4 Third generation computations: Current theoretical developments

4.1 Beyond electronic calculations: Anharmonicity

To go beyond the harmonic approximation is not straightforward, but it is essentialfor understanding hot media in which the temperature influences the band profilesand the intensities as a consequence of the anharmonicity of the potential energysurfaces. Starting from a static quantum chemical calculation, Born Oppenheimerdynamics can be used at two levels of approximation:

- either the vibrations are considered as independent , which takes into accountonly one-dimension anharmonicity;

- or the vibrations are considered as coupled.

As for electronic calculations, variational or perturbative treatments can be em-ployed, with similar pros and cons all linked to the size and nature of the basis set.A general option is to define a pool of vibrational states that can be populated atthe temperature considered and reachable via allowed IR transitions.

The problem of size, in fact the dimension of the basis of vibrational functionsso defined, becomes rapidly a bottleneck for the variational methods, especiallyin terms of computer time. For perturbation methods, it should be remindedthat possible bias can be introduced by the perturbative technique used in mostavailable codes, when accidental degeneracy occurs between vibrational levels,which is rather common when the size of the molecule and consequently the numberof vibrations increases. The solution is then to resort to degenerate perturbationtheory by solving variationally the manifold of quasi degenerate vibrational statesin a preliminary step (see for example the case of naphthalene in Cane et al. 2007or Pirali et al. 2009).

Using anharmonic parameters deduced from DFT results, it is possible to takethe participation of every populated state to the overall profile of each IR bandinto account. But as the number of transitions and of states to be consideredare quickly exploding with temperature, this type of calculation is so far limitedto small PAHS. The description of really hot bands has to be done by using astatistical Monte Carlo sampling for example. (G. Mulas et al. 2006a; Pirali et al.2009).


Anharmonic shifts Hot band profile

Fig. 4. Computational experiments on the absorption of naphthalene at fixed tempera-

tures. Left: calculated partial spectrum near 780 cm−1 at 500 K (adapted from Basire

et al. 2009); right: high resolution spectrum near 472 cm−1 at 300 K (adapted from

Pirali et al. 2009); experimental (top); calculated (bottom).

Another theoretical approach is actually proposed to calculate the finite tem-perature IR absorption spectrum of fully coupled anharmonic systems. The en-ergy levels are described by a second order Dunham expansion and a Wang Lan-dau Monte Carlo procedure is used for the calculation of the quantum densities ofstates; microcanonical simulations provide an absorption intensity function of boththe absorption wavelength and the internal energy of the molecule (Basire et al.2008; Basire et al. 2009; see also Basire in this volume). Here too, applicationsare presently limited to the smallest of the PAHs.

Selected results from both methods are given in Figure 4.

For larger systems, anharmonic effects can be obtained by Born-Oppenheimeror Car-Parrinello molecular dynamics (MD). Though MD methods cannot provideas detailed structures as the preceding methods, it can give the overall band evo-lution with temperature, and has the advantage of being able to deal with muchlarger systems efficiently. For even more effectiveness, it can be coupled to com-puter efficient calculations of the potential energy function, as a parametrized DFTbased on a tight binding approach such as SCC-DFTB (Self-Consistent ChargeDensity Functional Tight Binding (Porezag et al. 1995)); such an option used forSiPAH cations can be found in Joalland et al. (2010).


NeutralClosed shellσ unsaturation: NOπ hole: NO

ICH = VS

+

IonizedOpen shellσ unsaturation: NOπ hole: YES

NeutralOpen shellσ unsaturation: YESπ hole: NO

ICH = VS

+

IonizedClosed shellσ unsaturation: YESπ hole: NO

ICH = VSICH = VW

Fig. 5. Typical structures of neutral and ionized PAHs taken from the naphthalene series

(adapted from Pauzat et al. 2010).

4.2 Emission versus absorption spectra

The a priori construction of emission spectra that is the last step toward a directconfrontation of theoretical models to the UIR bands has only been done recently.Two classes of emission models have been proposed.

The thermal emission model describes the IR cooling of PAHs via transitionsfrom levels v to v− 1 in an emission cascade. The emission in a given vibrationalmode from a PAH of internal energy U is calculated as the average emission ofan oscillator connected to a thermal bath, the PAH itself being considered asa heat bath at the temperature T that corresponds to the average energy E ofthe molecule. This kind of model has been employed by Pech et al. (2002),using photophysical experimental data, as well as Pathak & Rastogi (2008) usingDFT B3LYP calculations (C10H8 → C96H24) for the initial adsorption spectra.It has to be noted that this method is efficient but rather approximate: indeed,this canonical formalism is good only when the excitation energy is much largerthan all the vibration modes considered, so that the average value of energy E isa sharply peaked function of T; it is a poor approximation in all other cases.

Mulas and co chose a micro-canonical formalism. This approach is based on theconservation of energy between emissions , which insures its validity whatever theexcitation energy considered. Though more difficult to implement than the thermalformalism, this option is more rigourous. Practically, the authors use a MonteCarlo technique on top of quantum chemical calculations by extending a precedingMonte Carlo model of photophysics to include rotational and anharmonic bandstructures and employing molecular parameters calculated by DFT (Mulas 1998;Mulas et al. 2006b).

4.3 Origin of the collapse of the CH intensity with ionization

Though the difference between the IR spectra of polycyclic aromatic hydrocarbons(PAH) and those of the corresponding positive ions (PAH+) is nowadays a welladmitted and documented fact, it remained puzzling and not clearly understoodfrom a theoretical point of view till very recently. Interpreting the origin of thesevariations has long been neglected, because the production of IR spectra of all sortsof PAHs was considered a matter of urgency in order to support the observationsand also because the tools available at the time were considered not reliable enoughfor such a delicate analysis.


Electronic states Level of theory

H

H

H

CH (2�)

CH+ (1�)

CH+ (3�)

Neutral PAH - Complete � system

Ionized PAH - Hole in � system

Ionized PAH - Hole in � system

0.6

0.8

1

1.2

1.4

1.6

1.8

0.95 1 1.05 1.1 1.15 1.2 1.25 1.3Dipole moment (Debye)

Internuclear distance (Angstrom)

B3LYPMRCISD

QMC

CH (2�)

CH+ (1�+)

CH+ (3�)

Lewis structures of CH Dipole moment functions

Fig. 6. Variation of the dipole moment of CH with the inter nuclear distance (adapted

from Pauzat et al. 2010).

When looking more precisely to the large body of data available on PAHs,such a behavior seems to be linked to species with a hole in the system, which isthe case of most classic ionized PAHs. A few demonstrative examples built fromnaphthalene are given hereafter in Figure 5 for a better understanding.

It is well known from vibrational analysis that the CH stretching vibration isa localized mode. Then, a local model of vibrator, namely, the CH fragment canrepresent the different types of PAHs as shown in Figure 6-left.

The study of the behavior of the dipole moment functions μ calculated withthe most elaborate methods available, state of the art ab-initio (MRCISD) andQuantum Monte Carlo (QMC), showed that the three model molecules exhibitvery different variations of their dipole moment with the CH distance (Fig. 6-right) that are consistent with strong/weak intensities of the CH vibrations in theneutral/ionized PAHs, the key point being the presence, or not, of a hole in theπ shell. This can be interpreted as the result of the distorsions of the electronicdensities, linked to different diabatic dissociation limits.

A topological analysis for the electronic densities (Pauzat et al. 2010) showsthat the collapse of the CH stretching is directly linked to the compensation be-tween the internal charge transfer contribution and the distortion of the electronicdensity within the CH bond, which occurs only in the case of CH+(1Σ+).

5 Conclusion

The PAH hypothesis has been a real source of motivation for several theoreticalgroups over the last two decades. This is a quasi text book example of creative in-teraction between space observation, laboratory experiments and theoretical com-putations. The following summary of the marking theoretical contributions to the


Year Landmarks Research groups1988 Poster IAU Symp. 135, Interstellar Dust Miller/DeFrees

naphthalene; anthracene (neutral/cation) MRI-IBM92-94 Effects of ionization (→ C16H10) Obs. Paris/ENS93 Anharmonicity and the 3.4 μm band ”95-97 Ionization versus dehydrogenation ”1996 Compilation of 13 PAHs (neutral/cation/anion) NASA Ames

→ C32H14 (ovalene)97 Perdeuterated small PAHs ”98 Effect of O/N substitution on Naphthalene ”99 PAHs with 5-membered rings ”99 Effect of double ionization Obs. Paris/ENS

Anthracene-Fe+ complex ”Effect of aliphatic substitution (3.4 μm region) ”

2000 PAHs (neutral/cation/anion/ multiple cations) NASA Ames→ C54H18 (circumcoronene)

01 Effects of perhydrogenation (3.4 μm region) Obs. Paris/ENSSpectra of Perhydrogenated PAHs NASA AmesLarge PAHs (neutral/protonated/cation) ”→ C96H24 (circumcircumcoronene)

02 Effects of non-regularity (neutral/cation) Obs. Paris/ENSPAHs with (4/5/7) membered ringsModel for IR emission (band profiles) U. Toulouse

03-05 PANHs (neutral/cation) NASA Ames04 Effect of deuteration ”05 Catacondended PAHs ( (neutral/cation) U. Gorakpur06 Pericondended PAHs ( (neutral/cation) ”

Far IR of PAHs: emission modeling Obs. Cagliari07 IR emission of PAHs dications ”

Fe+ PAH complexes U. Toulouse07-08 PAH IR emission modeling and size effects U. Gorakpur08 IR emission for PAHs with 10 to 96 carbons ”09 IR absorption of thermally excited naphthalene Obs. Cagliari

(anharmonic levels)IR spectra of Fe+/Mg+ PAH complexes NASA AmesMid IR of very large irregular PAHs ”C84H24 → C120H36

09-10 Anharmonic/temperature effects on IR spectra U. Paris XISiPAH+π complexes (anharmonicity) U. ToulouseFar IR of large to very large PAHs (→ 130 C) NASA AmesOrigin of the collapse of the 3.3 μm band U. Paris VIa

a Formerly Obs. Paris/ENS.

Fig. 7. Chronology of the main advances in theoretical studies of PAH infrared spectra.


knowledge of the infrared signatures of PAHs (Fig. 7) is certainly not exhaustive,the individual lonely contributions, except for the historical first one, not beingquoted but only those of full groups. Some of these groups, among which theNASA Ames and Toulouse University/Cagliari groups have the particularity andthe advantage to have an experimental group also dedicated to the PAH study,which allows them to constructive comparisons. These two groups are at the originof the two data bases available. During the last years, new young research groupshave entered the game and promoted real theoretical advances in particular forthe problems of anharmonicity and emission.

The PAH story is still going on!

References

Allamandola, L.J., Tielens, A.G., & Barker, J.R. 1985, ApJ, 290, L25

Amos, R.D., 1984, Chem. Phys. Lett., 108, 185

Bakes, E.L., Tielens, A.G., & Bauschlicher, C.W., 2001, ApJ, 556, 501

Basire, M., Parneix, P., & Calvo, F., 2008, J. Chem. Phys., 129, 081101

Basire, M., Parneix, P., Calvo, F., Pino, T., & Brechignac, Ph., 2009, J. Phys. Chem. A,113, 6947

Bauschlicher, C.W., & Langhoff, S.R., 1997, Spectrochim. Acta A, 53, 1225

Bauschlicher, C.W., Langhoff, S.R., Sandford, S.A., & Hudgins, D.H., 1997, J. Phys.Chem. A, 101, 2414

Bauschlicher, C.W., Hudgins, D.M., & Allamandola, L.J., 1999, Theor. Chem. Acc., 103,154

Bauschlicher, C.W., & Bakes, E.L., 2000, Chem. Phys., 262, 285

Bauschlicher, C.W., Hudgins, D.M., & Allamandola, L.J., 2001, Theor. Chem. Acc., 105,154

Bauschlicher, C.W., Peeters, E., & Allamandola, L.J., 2008, ApJ, 678, 316


Bauschlicher, C.W., & Ricca, A., 2009, ApJ, 698, 255

Bauschlicher, C.W., Boersma, C., Ricca, A., et al., 2010, ApJS, 189, 341

Becke, A.D., 1993, J. Chem. Phys., 98, 5648

Beegle, L.W., Wdowiak, T.J., & Harrison, J.G., 2001, Spectrochim. Acta A, 57, 737

Boersma, C., 2009, PhD, U. Groeningen

Cane, E., Miani, A., & Trombetti , A., 2007, J. Phys. Chem. A, 111, 8218

Cassam-Chenaı, P., Planet. Space Sci., 2002, 50, 871

DeFrees, D.J., Miller, M.D., Talbi, D., Pauzat, F., & Ellinger, Y., 1993, ApJ, 408, 530

Duley, W.W., & Williams, D.A., 1981, MNRAS, 196, 269

Ellinger, Y., Pauzat, F., & Lengsfield, B. H., 1999, J. Mol. Struc., 458, 203

Helgaker, T.U., Jensen, H.J.A., & Jorgensen, P., 1986, J. Chem. Phys., 84, 6280

Hudgins, D.H., Bauschlicher, C.W., & Allamandola, L.J., 2005, ApJ, 632, 316

Joalland, B., Simon, A., Marsden, C. J., & Joblin, C., 2009, A&A, 494, 969

Joalland, B., Rapacioli, M., Simon, A., et al., 2010, J. Phys. Chem. A, 114, 5846


Johnson, R.D., Irikura, K.K., Kacker, R.N., & Kessel, R., 2010, J. Chem. Theory Com-put., in press

Koch, W., & Holthausen, M.C., 2001, A Chemist Guide to Density Functional Theory,2nd ed. (Wiley-VCH, Weinheim, Germany)

Langhoff, S.R., 1996, J. Phys. Chem., 100, 2819

Langhoff, S.R., Bauschlicher, C.W., Hudgins, D.M., Sandford, S.A., & Allamandola, L.J.,1998, J. Phys. Chem. A, 102, 1632

Lee, C., Yang, W., & Parr, R.G., 1988, Phys. Rev. B, 37, 785

Lee, T.J., & Scuseria, G.E., 1995, In “Quantum mechanical electronic structure calcu-lations with chemical accuracy”, ed. S.R. Langhoff (Kluwer Academic Publishers:Dordrecht, The Netherlands)

Leger, A., & Puget, J.L., 1984, A&A, 137, L5

McLean, A.D., Lengsfield, B.H., Pacansky, J., & Ellinger, Y., 1985, J. Chem. Phys., 83,3567


Malloci, G., Joblin, C., & Mulas, G., 2007, A&A, 462, 627

Martin, J.M.L., El-Yazal, J., & Francois, J.P., 1994, J. Phys. Chem., 98, 4243

Merrick, J.P., Moran, D., & Radom, L., 2007, J. Phys. Chem. A, 111, 11683

Miller, M.D., & DeFrees, D.J., 1988, IAU Symp. 135, Interstellar Dust, Santa Clara

Mulas, G., 1998, A&A, 338, 243

Mulas, G., Malloci, G., & Benvenuti, P., 2003, A&A, 410, 639

Mulas, G., Malloci, G., Joblin, C., & Toublanc, D., 2006a, A&A, 456, 161

Mulas, G., Malloci, G., Joblin, C., & Toublanc, D., 2006b, A&A, 460, 93

Pathak, A., & Rastogi, S., 2007, Advances in Space Research, 40, 1620

Pathak, A., & Rastogi, S., 2007, Spectrochim. Acta Part A, 67, 898

Pathak, A., & Rastogi, S., 2008, A&A, 485, 735

Pauzat, F., Talbi, D., Miller, M.D., DeFrees, D.J., & Ellinger, Y., 1992, J. Phys. Chem.,96, 7882

Pauzat, F., Talbi, & Ellinger, Y., 1995, A&A, 293, 263

Pauzat, F., Talbi, & Ellinger, Y., 1997, A&A, 319, 318

Pauzat, F., Talbi, & Ellinger, Y., 1999, MNRAS, 304, 241

Pauzat, F., & Ellinger, Y., 2001, MNRAS, 324, 355

Pauzat, F., & Ellinger, Y., 2002, Chem. Phys., 280, 267

Pauzat, F., Pilme, J., Toulouse, J., & Ellinger, Y., 2010, J. Chem. Phys., 133, 054301


Peeters, E., Allamandola, L.J., Bauschlicher, C.W., et al., 2004, ApJ, 604, 252


Porezag, D., Frauenheim, T., Kohler, T., Seifert, G., & Kaschner, R., 1995, Phys. Rev.B, 51, 12947

Ricca, A., Bauschlicher, C.W., Mattioda, A.L., Boersma, C., & Allamandola, L.J., 2010,ApJ, 709, 42

Ricks, A.M., Douberly, G.E., & Duncan, M.A., 2009, ApJ, 702, 301


Scott, A., & Radom, L., 1996, J. Phys. Chem. A, 100, 16502

Simon, A., & Joblin, C., 2007, J. Phys. Chem., 111, 9745


Szabo, A., & Ostlund, N.S., 1982, Modern Quantum Chemistry (McGraw-Hill,New York)

Wakelam, V., & Herbst, E., 2008, ApJ, 680, 371

Yoshida, H., Ehara, A., & Matsuura, H., 2000, Chem. Phys. Lett., 325, 477


MODELING THE ANHARMONIC INFRARED EMISSIONSPECTRA OF PAHS: APPLICATION TO THE PYRENE

CATION

M. Basire1, P. Parneix1, T. Pino1, Ph. Brechignac1 and F. Calvo2

Abstract. The IR emission cascade from the pyrene cation due to abroad band optical excitation is simulated using kinetic Monte Carlo.Anharmonicities of the ground electronic state potential energy surfaceare taken into account in the transition energies, the microcanonicaldensities of states, and the rate of hydrogen loss through various sta-tistical theories. The emission spectral features of the “3.3”, “6.2” and“11.2” μm bands are computed for different blackbody temperatures.

1 Introduction

Unidentified Infrared Bands (UIBs) correspond to intense IR emission bands in the3–20 μm spectral range. According to the so-called PAH model (Puget & Leger1989; Allamandola et al. 1989); these IR bands, first detected in 1973 (Gillett et al.1973), are attributed to PAH emission bands, although no conclusive identificationof molecular carriers has been achieved yet. In the last years, a large numberof astrophysical objects has been observed in this IR spectral range and highresolution spectra are now available (Hony et al. 2001; Peeters et al. 2002; VanDiedenhoven et al. 2004; Draine & Li 2007). In the photophysics PAH model,the molecule is initially excited by absorption of a visible/UV photon, after whichthe intramolecular dynamics is driven by non adiabatic couplings. Subsequently,the PAH molecule is vibrationally heated in the ground electronic state and coolsdown by emission of several IR photons.

Up to now, only very few experimental data have been collected on the IRemission of neutral (Cherchneff & Barker 1989; Brenner & Barker 1992; Williams &Leone 1995; Cook et al. 1996; Cook & Saykally 1998) and cationic (Kim et al. 2001;

1 Institut des Sciences Moleculaires d’Orsay, CNRS UMR 8214, Universite Paris-Sud 11,Bat. 210, 91405 Orsay Cedex, France2 LASIM, Universite de Lyon and CNRS UMR 5579, 43 Bd. du 11 Novembre 1918,69622 Villeurbanne Cedex, France



Kim & Saykally 2002) PAHs after an initial excitation provided by a UV laser or anelectron gun. These experiments have found evidence that IR emission spectra areshifted with respect to absorption spectra, the vibrational shift being dependenton the amount of internal energy deposited in the molecule.

From the theoretical point of view, IR emission spectra have been mainlycomputed in the harmonic limit. In these simulations, the IR radiative cooling ofthe molecule is monitored using thermal approximations (Leger et al. 1989; Schutteet al. 1993) or a stochastic model (Gillespie 1978; Barker 1983; Mulas 1998; Joblinet al. 2002). More recently, in order to reproduce the shape of the spectral profile,anharmonicities were empirically introduced in the spectral simulation (Cook &Saykally 1998; Pech et al. 2002), using experimental data on the spectral shiftsand broadenings of the IR absoption spectra for neutral PAHs, as a function ofincreasing temperature (Joblin et al. 1995).

In the present article, we describe a novel theoretical approach to determinethe IR radiative cascade. In this method, anharmonicities calculated using first-principle methods and second-order perturbation theory are taken into account inall statistical quantities, namely the microcanonical densities of states used in thethermal populations and the dissociation rates, as well as the transition energiesthemselves.

2 Model and methods

We assume that the initial excitation energy, deposited as visible/UV light, israpidly converted into vibrational energy in the ground electronic state. Ourpurpose is then to model the (much slower) radiative decay of the molecule throughmultiple emission in the infrared domain. In absence of collisional relaxation (lowdensity medium), the internal energy remains constant between two successiveIR photon emissions, and the physical conditions are those of a microcanonicalensemble. If the time scale between two such emissions is long with respect tointramolecular vibrational redistribution (IVR), then all vibrational states at theavailable internal energy E are equiprobable.

Because the excitation energy is large with respect to the emitted IR pho-tons, the energy in the electronic ground state is expected to be initially ratherhigh, hence it is important to account for anharmonicities in the description oftransition energies, as well as for counting accurate densities of states (DOS’s).We have previously developed a computational procedure based on the Wang-Landau Monte Carlo method (Wang & Landau 2001) to determine the quantumdensity of vibrational states of fully coupled polyatomic systems. Briefly, themethod starts with a perturbative Dunham expansion of the vibrational energylevels in terms of harmonic wavelengths λ

(h)k and anharmonic coefficients χk�, as

obtained from quantum chemistry calculations. In the present case, all ingredientswere determined using density-functional theory, with the hybrid B3LYP potentialand the basis set 4-31G, of modest size because of computer limitations for thisrather large molecule. The quadratic force field was then sampled by performinga random walk in energy space, the microcanonical DOS being estimated with

M. Basire et al.: Modeling IR Emission Spectra 97

0 5 10Internal energy (eV)

0

Pro

babi

lity

0 10000 20000 30000 40000Tbb (K)

0

20

40

60

80

Dis

soci

atio

n ef

ficie

ncy

(%)

(a) (b)

Fig. 1. (a) Distribution of internal excitation energy P (E∗) at blackbody temperature

Tbb = 10 000 K; (b) dissociation efficiency versus blackbody temperature Tbb.

increasing accuracy over a broad energy range (Basire et al. 2008). Knowledge ofthe DOS also provides a very convenient way of calculating the infrared emissionspectrum at microcanonical equilibrium, by accumulating a two-dimensional his-togram Ie(ν; E) of emission frequency ν at total energy E along a multicanonicalMonte Carlo simulation (Basire et al. 2009, 2010). The energy-resolved spectrumis explicited as

Ie(ν; E) =1

N (ν; E)

∑MC steps

∑k

nkA(1→0)k δ

(hν − ΔE(k)

), (2.1)

where N (ν; E) is the number of entries in the histogram corresponding to theemission of a photon with energy hν, at total energy E. In this equation, nk is thevibrational level of mode k with harmonic frequency ν

(h)k , ΔE(k) is the transition

energy associated with the emission of a single photon of mode k, and nkA(1→0)k

is the associated oscillator strength, assuming here an harmonic approximation.In practice, the microcanonical spectra of the pyrene cation have been com-

puted using the above procedure for vibrational energies ranging from the groundstate level (Emin = 0) to Emax = 14 eV, with a resolution ΔE = 100 cm−1 in en-ergy and Δν = 0.2 cm−1 in emission frequency. Anharmonicities enter the aboveingredients through the densities of microcanonical states and thermal popula-tions, and in the transition energies.

Next, a complete IR emission cascade is simulated using a kinetic Monte Carlo(kMC) model (Mulas 1998). At any given energy E, several photon emissionprocesses compete with each other and with molecular dissociation. For simplicity,we only include the lowest (most probable) dissociation channel, namely hydrogenloss, for which the dissociation rate kdiss(E) is accurately modeled using phasespace theory (Pino et al. 2007). Concerning radiative relaxation, the emission ratek(ν; E) for a photon hν at internal energy E is simply given by Ie(ν; E). Startingat time t = 0 with an excess energy E(t = 0) = E∗, all possible relaxation eventsare enumerated, including emission of a single photon from mode k with nk > 0,


3.2 3.40In

tens

ity (

arb.

uni

t)

6.2 6.4 6.6 6.80

11.2 11.4 11.60

3.2 3.4Wavelength (μm)

0Inte

nsity

(ar

b. u

nit)

6.2 6.4 6.6 6.8Wavelength (μm)

011.2 11.4 11.6Wavelength (μm)

0

(a) (b) (c)

(d) (e) (f)

Fig. 2. Simulated emission spectra of the pyrene cation. In panels (a), (b), and (c), the

blackbody temperature is Tbb = 6000 K; in panels (d), (e), (f), Tbb = 2 × 104 K. The

spectral ranges considered are 3.1–3.5 μm in panels (a) and (d), 6.2–6.8 μm in panels (b)

and (e), and 11.1–11.7 μm in panels (c) and (f).

together with hydrogen dissociation, and one of them is randomly drawn with aprobability proportional to its rate. The time and internal energy are updated,and the process is repeated until dissociation occurs or if time exceeds 10 s. 107

such kMC trajectories were performed in order to build reasonably smooth IRemission spectra.

In this article, only large band black-body excitations are considered. Theabsorption cross-section σ(E) was adapted from experimental data (Salama &Allamandola 1992; Biennier et al. 2004) in the low energy range (E ≤ 4 eV) andfrom theoretical works (Malloci et al. 2004) for 4 < E ≤ 12 eV. The initial elec-tronic excitation E∗ under the blackbody radiation flux Φ(E) is randomly drawn atthe beginning of each kMC trajectory from the probability P (E∗) ∝ σ(E∗)Φ(E∗),after suitable normalization. As an illustration, the visible/UV absorption prob-ability for a black-body temperature Tbb = 6000 K is displayed in Figure 1a. Atthis temperature, absorption mainly occurs near 6 eV due to large cross-sections ofexcited electronic states in this energy range. The dissociation efficiency is shownin Figure 1b versus Tbb. Dissociation is found to play a role in the photodynam-ics of the pyrene cation only when Tbb exceeds about 1000 K. In the temperaturerange considered here, the dissociation efficiency is always lower than 40%. As willbe seen below, the dissociation involves hydrogen loss, which significantly altersthe infrared emission spectral features.


3 Results and discussion

For the sake of brevity, we focus on the evolution of IR emission spectral featuresas a function of the blackbody temperature associated to the stellar radiation flux.Three vibrational bands have been considered, namely “3.3” μm (C-H stretchings),“6.2” μm (C-C stretchings) and “11.2” μm (C-H out-of-plane bendings), owing totheir astrophysical relevance. A more detailed analysis of the IR emission bandswill be given in a forthcoming article. In Figure 2, the emission spectra for thesethree bands are shown for Tbb = 6000 K and 2 × 104 K.

The calculated emission spectra are broad, continuous, and strongly asym-metric, especially for the two softer bands where long wings on the red side areconsistent with observations (Van Diedenhoven et al. 2004; Peeters et al. 2002).This spectral profile is in agreement with previous calculations in which spectralshifts and broadenings due to anharmonicity were extracted from experimentaldata (Pech et al. 2002). The asymmetry also increases with blackbody tempera-ture, and the profile changes significantly in the case of the “6.2” μm band. Allthese effects originate from anharmonicities of the ground state potential energysurface.

On a more quantitative level, the spectral shift 〈λ〉 and the spectral widthΔλ have been computed as a function of Tbb using the first and second momentsof the IR emission spectrum, respectively. The variations of 〈λ〉 and Δλ withTbb are represented in Figure 3 for the same three vibrational bands depicted inFigure 2. Comparing 〈λ〉 to the harmonic wavelength λ

(h)k , the “11.2” band turns

out to be less sensitive to anharmonicities than the two others bands. At veryhigh blackbody temperatures, both the shift and width tend to constant valuesfor the three modes. This is a consequence of hydrogen dissociation, which at highenergies becomes an important channel very soon after excitation (see Fig. 1), thusquenching IR emission due to the concomitant strong evaporative cooling.

Finally, we note that the three simulated spectral widths obtained here forcationic pyrene are of the same order of magnitude as those observed in the ma-jority of interstellar sources (Van Diedenhoven et al. 2004; Peeters et al. 2002).For the three bands considered so far (in decreasing wavelength), the observedwidths are approximately equal to 0.20 μm, 0.13 μm, and 0.04 μm, respectively,in agreement with Figure 3.

The observed spectral features seem rather unsensitive to the astrophysicalobjects in the class A (Peeters et al. 2002; Van Diedenhoven et al. 2004; Gallianoet al. 2008). The observed IR band broadening is the consequence of the PAHsize distribution. However, as demonstrated by the present calculations limitedto cationic pyrene, the asymmetric spectral profiles are well reproduced whenthe PES anharmonicity is properly incorporated, especially for the “11.2” μmband. The observed IR emission spectra could then be strongly influenced by thevariety in the intrinsic spectral width of individual PAHs resulting from differentanharmonicities. The present computational protocol offers a convenient way toaddress this issue for larger molecules on a quantitative footing. Work is currentlyin progress to analyse the corresponding size effects on IR emission spectroscopy.


10000 20000 300006.44

6.48

6.52

⟨λ⟩ (

μm)

10000 20000 300000.08

0.12

0.16

0.2

Δλ (

μm)

10000 20000 3000011.28

11.32

11.36

⟨λ⟩ (

μm)

10000 20000 300000.08

0.12

0.16

0.2

Δλ (

μm)

10000 20000 30000Tbb (K)

3.23

3.24

3.25

⟨λ⟩ (

μm)

10000 20000 30000Tbb (K)

0.03

0.04

0.05

0.06Δλ

(μm

)

(a)

(b)

(c)

(d)

(e)

(f)

Fig. 3. Variations of the spectral shift 〈λ〉 (left panels) and width Δλ (right panels) with

increasing blackbody temperature, for the same three vibrational bands of the pyrene

cation depicted in Figure 2. (a) and (d): 11.10 μm; (b) and (e): 6.25 μm; (c) and (f):

3.10 μm.

References

Allamandola, L.J., Tielens, A.G.G.M., & Barker, J.R., 1989, ApJS, 71, 733

Barker, J.R., 1983, Chem. Phys., 77, 301

Basire, B., Parneix, P., & Calvo, F., 2008, J. Chem. Phys., 129, 0811019

Basire, B., Parneix, P., Calvo, F., Pino, T., & Brechignac, Ph., 2009, J. Phys. Chem. A,113, 6947

Basire, B., Parneix, P., & Calvo, F., 2010, J. Phys. Chem. A, 114, 3139

Biennier, L., Salama, F., Gupta, M., & O’Keefe, A., 2004, Chem. Phys. Lett., 387, 287

Brenner, J., & Barker, J.R., 1992, ApJ, 388, L39

Cherchneff, I., & Barker, J.R., 1989, ApJ, 341, L21

Cook, D.J., Schlemmer, S., Balucani, N., et al., 1996, Nature, 380, 227

Cook, D.J., & Saykally, R.J., 1998, A&A, 493, 793

Draine, B.T., & Li, A.G., 2001, ApJ, 551, 807

Draine, B.T., & Li, A.G., 2007, ApJ, 657, 810



Gillespie, D.T., 1978, J. Comput. Phys., 28, 395

Gillett, F.C., Forrest, W.J., & Merrill, K.M., 1973, ApJ, 183, 87

Hony, S., Van Kerckoven, C., Peeters, E., et al., 2001, A&A, 370, 1030

Joblin, C., Boissel, P., Leger, A., & Defourneau, D., 1995, A&A, 299, 835

Joblin, C., Toublanc, D., Boissel, P., & Tielens, A.G.G.M., 2002, Mol. Phys., 100, 3595

Kim, H.S., Wagner, D.R., & Saykally, R.J., 2001, Phys. Rev. Lett., 86, 5691

Kim, H.S., & Saykally, R.J., 2002, ApJS, 143, 455

Leger, A., D’Hendecourt, L., & Defourneau, D., 1989, A&A, 216, 148

Malloci, G., Mulas, G., & Joblin, C., 2004, A&A, 426, 105

Mulas, G., 1998, A&A, 338, 243


Pino, T., Parneix, P., Calvo, F., & Brechignac, Ph., 2007, J. Phys. Chem. A, 111, 4450

Peeters, E., Hony, S., Van Kerckoven, C., et al., 2002, A&A, 390, 1089

Puget, J.L., & Leger, 1989, ARA&A, 27, 161

Salama, F., & Allamandola, L.J., 1992, Nature, 358, 42

Schutte, W.A., Tielens, A.G.G.M., & Allamandola, L.J., 1993, ApJ, 415, 397

Van Diedenhoven, B., Peeters, E., Van Kerckoven, C., et al., 2004, ApJ, 611, 928

Wang, F., & Landau, D.P., 2001, Phys. Rev. Lett., 86, 2050

Williams, R.M., & Leone, S.R., 1995, ApJ, 443, 675


LABORATORY SPECTROSCOPY OF PROTONATED PAHMOLECULES RELEVANT FOR INTERSTELLAR CHEMISTRY

O. Dopfer1

Abstract. In this contribution, we summarize the recent progress madein recording laboratory infrared (IR) spectra of protonated polycyclicaromatic hydrocarbon molecules (H+PAH) in the gas phase. The IRspectra of a large variety of H+PAH species ranging from benzene tocoronene have been obtained by various variants of photodissociationspectroscopy. The employed techniques include single-photon IR pho-todissociation (IRPD) of tagged H+PAH ions and IR multiple-photondissociation (IRMPD) of bare H+PAH ions. The comparison of the lab-oratory IR spectra with astronomical spectra supports the hypothesisthat H+PAH ions are possible carriers of the unidentified IR emission(UIR) bands. Moreover, the spectra provide detailed information aboutthe geometric and electronic structure as well as the chemical reactivityand stability of these fundamental hydrocarbon ions.

1 Introduction

The unidentified infrared (UIR) emission bands are observed in a great variety ofgalactic and extragalactic sources. They are currently attributed to IR fluorescenceof UV-excited polycyclic aromatic hydrocarbon (PAH) molecules and their morecomplex derivatives, such as ionic PAH, (de-)hydrogenated and methyl-substitutedPAH, PAH clusters, nitrogen- and silicon-containing PAH, and PAH-metal ionadducts (Leger & Puget 1984; Allamandola et al. 1985; Tielens 2008). Recently,also protonated PAH molecules, H+PAH, have been suggested because laboratoryexperiments demonstrated that they can readily be formed through efficient pro-tonation of neutral PAH or attachment of H-atoms to PAH+ radical cations (Snowet al. 1998). In addition, their calculated IR spectra identify H+PAH as promisingcandidates for UIR band carriers (Hudgins et al. 2001).

In order to test the H+PAH hypothesis, laboratory IR spectra of H+PAHare required for comparison with astronomical UIR spectra. Such spectra have,

1 Institut fur Optik und Atomare Physik, Technische Universitat Berlin, Hardenbergstrasse36, 10623 Berlin, Germany; e-mail: [email protected]



however, been lacking until recently (Lorenz et al. 2007), largely due to exper-imental challenges involved in the generation of high abundances sufficient forspectroscopic interrogation. Recent progress in the development of efficient ionsources and sensitive IR spectroscopic techniques have allowed to obtain the firstIR spectra of size-selected protonated aromatic molecules and their weakly-boundclusters in the gas phase by coupling tandem mass spectrometry with resonantvibrational photodissociation spectroscopy (Solca & Dopfer 2001; Solca & Dopfer2003a). The application of these techniques to protonated aromatic molecules hasbeen reviewed previously (Dopfer 2006). In this contribution, we summarize theresults obtained in the past decade for a large variety of H+PAH ions, rangingfrom protonated benzene to protonated coronene.

2 Experimental techniques

Several experimental approaches have been used to generate isolated and micro-solvated H+PAH cations in the gas phase for spectroscopic interrogation. Ion andcluster ion generation has been accomplished in an electron-impact or discharge-driven supersonic plasma expansion. This approach was used to produce pro-tonated benzene and naphthalene, C6H+

7 and C10H+9 , and their weakly-bound

clusters with ligands L = Ar, N2, CH4, and H2O (Solca & Dopfer 2002; Solca &Dopfer 2003b; Douberly et al. 2008; Ricks et al. 2009). Alternatively, isolatedH+PAH can be generated by chemical ionization of PAH in an ion cyclotron reso-nance (ICR) mass spectrometer using protonating agents, such as CH+

5 or C2H+5 ,

as successfully demonstrated for C6H+7 and C10H+

9 (Jones et al. 2003; Dopferet al. 2005; Lorenz et al. 2007). Larger PAH molecules have low vapour pressureand a feasible option to generate sufficient abundances of their protonated speciesis electrospray ionization. This technique has been applied to characterize largerH+PAH ions (Knorke et al. 2009), ranging from protonated anthracene, C14H+

11,to protonated coronene, C24H+

13.Two major spectroscopic strategies have successfully been employed to obtain

IR spectra of isolated and microsolvated H+PAH ions. The first technique couplesmodern low-intensity optical parametric oscillator (OPO) laser systems, typicallyoperating in the 1000–4000 cm−1 range, with tandem mass spectrometers to recordsingle-photon IRPD spectra of cold H+PAH-Lm cluster ions by monitoring theloss of weakly-bound ligands L (Solca & Dopfer 2002; Solca & Dopfer 2003b;Douberly et al. 2008; Ricks et al. 2009). The low intensities of OPO lasersare usually insufficient to drive multiple-photon processes. The influence of theligands on the IR spectrum of the central H+PAH ion can usually be neglectedfor weakly-bound ligands (e.g., Ar), i.e. the IRPD spectra of the tagged speciescorrespond to a very good approximation (to within a few cm−1) to the IR spectraof the bare ion. Moreover, the clusters are cold as they are produced in molecularbeams. As a result of the low internal temperature (typically well below 50 K)and the single-photon dissociation process, IRPD spectra display high resolutionand sharp transitions with widths of the order of 5 cm−1. Moreover, they displayhigh sensitivity, i.e. even transitions with low IR activity can readily be detected.

O. Dopfer: Laboratory IR Spectra of Protonated PAH 105

The IRPD approach has been applied to obtain the first IR spectrum of C6H+7 in

the gas phase (Solca & Dopfer 2002).The second spectroscopic strategy couples high-intensity free electron laser

(FEL) sources (FELIX and CLIO), operating in the complementary 50–2500 cm−1

range, with ion cyclotron resonance (ICR) mass spectrometry to drive IRMPD ofisolated, strongly-bound H+PAH by monitoring either H-atom or H2 loss (Joneset al. 2003; Dopfer et al. 2005; Lorenz et al. 2007; Zhao et al. 2009; Knorkeet al. 2009). In general, the spectral resolution observed in the IRMPD spectra(circa 30 cm−1) is lower than in IRPD spectra due to the finite bandwidth of theFEL, the broader rotational contour (as the ions probed in the ICR cell are atroom temperature), spectral congestion due to overlapping transitions, and spec-tral broadening arising from the multiple-photon character of the IRMPD process.Typically, several tens of IR photons are required to drive the IRMPD process dueto the large dissociation energies for H or H2 elimination from H+PAH. As a fur-ther consequence of the IRMPD mechanism, weakly IR active transitions withcross sections below a certain threshold are not observed. The IRMPD approachhas been utilized to obtain the first IR spectrum of an isolated H+PAH, namelyC10H+

9 (Lorenz et al. 2007). Figure 1 compares the IRMPD spectra of a variety ofH+PAH obtained so far, including protonated benzene, naphthalene, anthracene,tetracene, pentacene, azulene, perylene, pyrene, coronene, acenaphthene, and ace-naphthylene. While this contribution focuses on the general appearance of theIR spectra and the comparsion with an UIR spectrum representative of a highly-ionized region of the interstellar medium (Tielens 2008; Fig. 1), a full accountof the potential energy surfaces with structural and vibrational analyses is givenelsewhere (Zhao et al. 2009; Zhao 2010). Significantly, apart from benzene H+,the IRMPD spectra in Figure 1 correspond to the first spectroscopic data of theseH+PAH in the gas phase, providing very valuable information about the geometricand electronic structures as well as the chemical reactivity and stability of thesefundamental ions. For an illustration of the information content of these spectra,the reader is referred to a recent analysis of the azuleneH+ spectrum using quan-tum chemical calculations (Zhao et al. 2009). A similar analysis was performedfor all H+PAH spectra in Figure 1 (Zhao 2010) and will be published elsewhere.

3 Results and discussion

Initial IRPD spectra obtained for C6H+7 in the CH stretch range using the tagging

technique (Solca & Dopfer 2002; Solca & Dopfer 2003b) demonstrated for the firsttime spectroscopically that the proton is attached to a C-atom (σ-complex) ratherthan the π electron system of the aromatic ring (π-complex), a conclusion that alsoholds for larger H+PAH. In addition, these spectra do not show coincidences withprominent UIR bands. Subsequent IRMPD spectra reported for isolated C6H+

7

in the fingerprint range (Fig. 1) using CLIO and monitoring H2 loss (Jones et al.2003; Dopfer et al. 2005) and higher-resolution IRPD spectra of Ar-tagged C6H+

7

ions (Douberly et al. 2008) confirm these initial conclusions.


600 1000 1400 1800

benzeneH+

naphthaleneH+

anthraceneH+

tetraceneH+

pentaceneH+

azuleneH+

peryleneH+

pyreneH+

coroneneH+

acenaphtheneH+

wavenumber (cm–1)

acenaphthyleneH+

600 800 1000 1200 1400 1600 1800 2000

Cor (Ar matrix)

Cor+ (Ar matrix)

CorH+ (IRMPD)

Cor+ (IRMPD)

UIR

6.2

6.9

7.6

7.8

8.6

11.0

11.2

12.7

13.6

14.2

16.4

wavenumber (cm–1)

Fig. 1. Left: IRMPD spectra of a variety of H+PAH ions (Dopfer 2005; Lorenz et al.

2007; Dopfer 2008; Knorke et al. 2009; Zhao et al. 2009); the spectra of benzene H+

and naphthalene H+ were recorded using CLIO, whereas all other spectra were obtained

using FELIX; bands marked by asterisks are largely suppressed due to reduced laser

power. Right: IR spectra of coronene, coronene + and coronene H+ compared to an UIR

spectrum (Knorke et al. 2009). The dashed lines indicate positions of UIR bands in μm.

Next, IRMPD spectra were recorded for C10H+9 , the smallest H+PAH, using

CLIO and the major conclusions can be summarized as follows (Lorenz et al.2007). First, the IRMPD spectra of C6H+

7 and C10H+9 are very different (Fig. 1),

implying a strong dependence of the IR spectra of H+PAH on the number ofaromatic rings, at least for small PAH. Second, the IRMPD spectrum of C10H+

9

demonstrates striking coincidences with the UIR spectrum. Both, the positionsand IR intensities of the C10H+

9 bands match the astronomical spectrum muchbetter than the IR spectra of neutral and ionized naphthalene. These importantconclusions were later confirmed by IRPD spectra of C10H+

9 -Ar (Ricks et al. 2009).

The close match of the C10H+9 and UIR spectra called for laboratory IR spectra

of larger H+PAH, in order to follow their evolution as a function of PAH size.This was also of importance in the light of astrochemical models, which suggestPAH molecules consisting of 20–80 C-atoms to be photochemically most stablein interstellar clouds (Tielens 2008). To this end, IRMPD spectra of a variety oflarger H+PAH species (Fig. 1) have been recorded using FELIX (Dopfer 2008).

O. Dopfer: Laboratory IR Spectra of Protonated PAH 107

Figure 1 compares the IRMPD spectra of linear catacondensed H+PAH fea-turing n aromatic rings with n = 1–5 (Knorke et al. 2009). The n = 1 and n = 2spectra are significantly different in appearance, and both differ from those forn = 3–5. However, the IRMPD spectra for n = 3–5 are rather similar, suggest-ing convergence of the spectral evolution at around n = 3. Comparison of then = 3–5 spectra with an UIR spectrum representative of a highly-ionized regionof the interstellar medium (Tielens 2008; Fig. 1) reveals good correspondence fortransitions assigned to in-plane CH bend and CC stretch modes. In addition, thehighest-frequency CC stretch mode at around 1600 cm−1 approaches the 6.2 μmUIR feature with increasing n. The intense, lower-frequency CC stretch modesat 1450 cm−1 are characteristic of all linear catacondensed H+PAH and corre-late with the 6.9 μm UIR feature, which is however only weakly observed in theastronomical spectrum. Similar conclusions apply to the 1500 cm−1 transition.

Not only are linear catacondensed H+PAH with n = 1–5 probably too smallto be photostable, they are also less stable than pericondensed two-dimensionalH+PAH. To this end, we recorded also the IRMPD spectrum of protonated coronene(C24H+

13), a pericondensed PAH, which is often considered as the prototypical andrepresentative PAH molecule present in the interstellar medium. Figure 1 com-pares the available experimental IR spectra for different degrees of ionization andprotonation of coronene with the UIR spectrum (Knorke et al. 2009). Comparisonof the spectra of coronene and coronene+ recorded in an Ar matrix demonstratesthe significant impact of ionization on the IR intensities. The coronene+ spectrumrecorded in the cold Ar matrix and via IRMPD in the gas phase are similar butillustrate the typical redshift and broadening of bands induced by the IRMPD pro-cess. The IRMPD spectrum of coronene+ and coroneneH+ are again similar. Theredshifts and broadening effects due to IRMPD are, however, less pronounced forthe protonated species due to its lower dissociation energy. Moreover, the bandsin the 1100–1200 cm−1 range display higher IR activity. Significantly, the intenseCC stretch band at 1600 cm−1 (6.25 μm) occurs at a higher frequency than forthe radical cation and closely approaches the 6.2 μm band observed in the UIRspectrum.

As the UIR bands are caused by IR fluorescence of vibrationally hot species,the vibrational transitions observed are shifted in frequency from the frequen-cies of fundamental transitions due to vibrational anharmonicities. Similarly, themultiple-photon nature of the IRMPD process also shifts the frequencies to lowervalues in comparison to fundamental transitions, again via the effect of anhar-monicities. Thus, it is indeed meaningful to compare IRMPD spectra with theUIR bands, as both processes and the resulting IR spectra are affected by thesame type of vibrational anharmonicity (Oomens et al. 2003). In view of thisargument, the good agreement between the IRMPD spectrum of coroneneH+ andthe UIR spectrum in Figure 1 provides for the first time compelling experimentalevidence for the hypothesis that H+PAH indeed contribute to the UIR spectrum.Future efforts include recording IR spectra of even larger H+PAH ions in orderto follow the spectral evolution as a function of PAH size. In addition, electronicspectra of H+PAH will be recorded, as they are promising candidates for carriers


of the diffuse interstellar bands (Pathak & Sarre 2008; see also Hammonds et al.,elsewhere in this volume). First encouraging results have been obtained for C10H+

9

(Alata et al. 2010), which strongly absorbs in the visible range, supporting thehypothesis that larger H+PAH ions may indeed be carriers of some of the diffuseinterstellar bands.

This work was supported by TU Berlin, the Deutsche Forschungsgemeinschaft (DO 729/2;729/3), and the European Community - Research Infrastructure Action under the FP6 Struc-turing the European Research Area Program (through the integrated infrastructure initiativeIntegrating Activity on Synchrotron and Free Electron Laser Science).

References


Alata, I., Omidyan, R., Broquier, M., et al., 2010, Phys. Chem. Chem. Phys., 12, 14456

Dopfer, O., Solca, N., Lemaire, J., et al., 2005, J. Phys. Chem. A, 109, 7881

Dopfer, O., 2006, J. Phys. Org. Chem., 19, 540

Dopfer, O., 2008, report for project FELIX-030

Douberly, G.E., Ricks, A.M., Schleyer, P.V.R., & Duncan, M.A., 2008, J. Phys. Chem.A, 112, 4869

Hudgins, D.M., Bauschlicher, C.W., & Allamandola, L.J., 2001, Spectrochim. Acta PartA, 57, 907

Jones, W., Boissel, P., Chiavarino, B., et al., 2003, Angew. Chem. Int. Ed., 42, 2057

Knorke, H., Langer, J., Oomens, J., & Dopfer, O., 2009, ApJ, 706, L66

Leger, A., & Puget, J.L., A&A, 137, L5

Lorenz, U.J., Solca, N., Lemaire, J., Maitre, P., & Dopfer, O., 2007, Angew. Chem. Int.Ed., 46, 6714

Pathak, A., & Sarre, P.J., 2008, MNRAS, 391, L10

Ricks, A.M., Douberly, G.E., & Duncan, M.A., 2009, ApJ, 702, 301

Snow, T.L., Page, L.V., Keheyan, Y., & Bierbaum, V.M., 1998, Nature, 391, 259

Solca, N., & Dopfer, O., 2001, Chem. Phys. Lett., 342, 191

Solca, N., & Dopfer, O., 2002, Angew. Chem. Int. Ed., 41, 3628

Solca, N. & Dopfer, O., 2003, J. Am. Chem. Soc., 125, 1421

Solca, N., & Dopfer, O., 2003, Chem. Eur. J., 9, 3154


Oomens, J., Tielens, A.G.G.M., Sartakov, B.G., von Helden, G., & Meijer, G., 2003,ApJ, 591, 968

Zhao, D., Langer, J., Oomens, J., & Dopfer, O., 2009, J. Chem. Phys., 131, 184307

Zhao, D., 2010, Diploma Thesis, TU Berlin, Germany


THE NASA AMES PAH IR SPECTROSCOPIC DATABASEAND THE FAR-IR

C. Boersma1, L.J. Allamandola1, C.W. Bauschlicher, Jr.2, A. Ricca2,J. Cami3, E. Peeters3, F. Sanchez de Armas4, G. Puerta Saborido4,

A.L. Mattioda1 and D.M. Hudgins5

Abstract. Polycyclic Aromatic Hydrocarbons (PAHs) are widespreadacross the Universe and influence many stages of the Galactic lifecycle.The presence of PAHs has been well established and the rich mid-IRPAH spectrum is now commonly used as a probe into (inter)stellar envi-ronments. The NASA Ames PAH IR Spectroscopic Database has beenkey to test and refine the “PAH hypothesis”. This database is a largecoherent set (>600 spectra) of laboratory measured and DFT computedinfrared spectra of PAHs from C10H8 to C130H28 and has been madeavailable on the web at (http://www.astrochem.org/pahdb). With anew spectral window opening up; the far-IR, the study of PAH far-IRspectra and the quest for identifying a unique member of the interstel-lar PAH family has begun. To guide this research, the far-IR (>20 μm)spectra of different sets of PAHs are investigated using the NASA AmesPAH IR Spectroscopic Database. These sets explore the influence ofsize, shape, charge and composition on the far-IR PAH spectrum. Thefar-IR is also the domain of the so-called “drumhead” modes and othermolecular vibrations involving low order bending vibrations of the car-bon skeleton as a whole. As with drums, these are molecule and shapespecific and promise to be a key diagnostic for specific PAHs. Here, thesensitivity of these “drumhead” modes to size and shape is assessed bycomparing the frequencies of the lowest drumhead modes of a familyof circular shaped (the coronene “family”) and rhombus shaped (thepyrene “family”) PAH molecules. From this study, some consequencesfor an observing strategy are drawn.

1 NASA Ames Research Center, MS 245-6, Moffett Field, CA 94035, USAe-mail: [email protected] NASA Ames Research Center, MS 230-3, Moffett Field, CA 94035, USA3 Department of Physics and Astronomy, PAB 213, The University of Western Ontario,London, ON N6A 3K7, Canada4 SETI Institute, 515 N. Whisman Road, Mountain View, CA 94043, USA5 NASA Headquarters, MS 3Y28, 300 E St. SW, Washington, DC 20546, USA



number of carbon atoms

number

of spe

cies

10 130100703020

50

0

100

200

150

50

`pure’nitrogenoxygen

magnesium & iron

composition

neutral42%

+44%

++3%

+++2%

-9%

other3%

oxygen19%

nitrogen21%

`pure’58%

Fig. 1. Left: size distribution, in terms of the number of carbon atoms, of the PAHs

in the computational database by composition; “pure” PAHs contain only carbon and

hydrogen, nitrogen/oxygen/magnesium & iron refer to PAHs containing these elements

as well. Middle: pie chart showing the breakdown by composition. Right: pie chart

showing the charge distribution (Bauschlicher et al. 2010).

1 Introduction

The astronomical emission features formerly known as the unidentified infraredbands are now commonly ascribed to polycyclic aromatic hydrocarbons (PAHs).The spectra from laboratory experiments and computational modeling done atthe NASA Ames Research Center to test and refine the “PAH hypothesis” havebeen assembled in a spectroscopic database. Here a few highlights are given from astudy of PAH far-IR spectra that utilizes this database (Boersma et al. submitted).This study, that complements that by Mulas et al. (2006) utilizing the Cagliaridatabase of theoretical PAH spectra (Malloci et al. 2007), should prove useful tocompare with data obtained by e.g., ESA’s Herschel satellite, the Atacama LargeMillimeter Array (ALMA), and NASA’s Stratospheric Observatory For InfraredAstronomy (SOFIA).

2 The NASA ames PAH IR spectroscopic database

The NASA Ames PAH IR Spectroscopic Database now contains over 600 PAHspectra, spanning 2–2000 μm, and is available at www.astrochem.org/pahdb.Tools that allow the user to: 1) interrogate the database; 2) plot, compare andco-add spectra, and 3) adjust parameters such as bandwidth etc. have also beenmade available online. Additionally, the database can be downloaded along witha suite of IDL tools (the AmesPAHdbIDLSuite).

Here the computational data, currently at version 1.11, are utilized and aredescribed in detail by Bauschlicher et al. (2010). The experimental data in thedatabase will be described by Mattioda et al. (in preparation). Figure 1 shows thedistribution of the computational database, broken down by size and composition.Figure 1 also summarizes the charge and PAH type distributions in two pie charts.

Due to improvements in computational methods and in computing power, it isnow possible to compute the spectra of PAHs containing more than one hundred

C. Boersma et al.: The Ames PAH Database and the far-IR 111

carbon atoms in reasonable amounts of time. When these new sets of spectrabecome available, they will be added to the database and the version numberwill be updated accordingly. One promising application is to use the database todirectly fit astronomical spectra (see Cami, elsewhere in this volume).

3 PAH far-IR spectroscopy – some highlights

Laboratory experiments and DFT computations show that PAH molecules havebands that span the far-IR from 20 to 1000 μm and, unlike the mid-IR, wherefundamental vibrational frequencies are determined largely by vibrations involvingthe chemical subgroups and specific bonds which make up the molecule, theselonger wavelength transitions originate from vibrations of the entire molecule.Thus, while the mid-IR spectra of PAHs resemble one another since all PAHs arepart of the same chemical class, beyond about 15 μm band positions can dependon overall size and structure.

3.1 PAH classes

The influence of PAH shape, charge, composition and size is studied in Figure 2.Shape: Comparing the effect of structural modifications on neutral circumcir-

cumcoronene (C96H24) show that with increasing irregularity, far-IR PAH spectratend to get richer in features.

Charge: The far-IR spectra of C78H22 in the –1, 0, +, +2 and +3 charge statesshows that varying charge does not have a strong influence on band positions.However, relative band intensities show some differences.

PAHNs: The far-IR spectra of C96H+24 with 8 single nitrogen containing iso-

mers shows that band positions and absolute intensities are hardly affected by theposition of the nitrogen atom in the hexagonal network.

Size: The far-IR spectra of ten compact neutral PAHs of increasing size showthat the spectra tend to get richer in features and extend further into the far-IRas the molecules get larger.

3.2 PAH “drumhead” frequencies

Figure 3 presents the lowest vibrational mode of four members of both the coroneneand pyrene “families”. The plot demonstrates the shift to lower frequencies of theso-called “drumhead” mode as molecular size increases. The vibrational frequen-cies are well fitted by:

ν = 600(

10−15 [cm2]A

)[cm−1], (3.1)

claiming an inverse dependence of frequency on PAH area (A) for the compactPAHs and only a weak dependence on molecular geometry; circular versus rhombusshaped.


Fig. 2. Series of far-IR PAH spectra (20 – 1000 μm) showing, from top-left-to-bottom-

right, the effect of shape, charge, nitrogen substitution, and size. Bands have been given

Lorentzian profiles with a FWHM of 6 cm−1. See Boersma et al. (submitted) for a

detailed discussion.

Considering PAH molecules as a solid plate, this is perhaps not that surprising.The classic solution for the frequency of vibration of “free” plates is given by


Fig. 3. Left: comparison of the predicted (0, 1) frequencies of a solid graphene plate,

Equation (3.2), with those found in the database for the coronene (squares) and pyrene

(triangles) “families”. Right: members of the coronene and pyrene “families”. From

Boersma et al. (submitted).

(Meirovitch 1997):

ν =π

2c

1A

√D

ρh· (n2 + m2), (3.2)

with c the speed of light, A the surface area, ρ the density, h the plate thicknessand D the rigidity. The modes are characterized by m and n; the number of nodesalong both plate-axes. Because of fundamental and commercial interest, there isa rich literature on the mechanical properties of graphene and carbon nano-tubes.The bending rigidity of a graphene sheet has been calculated to be 0.8 – 1.5 eV,depending on the method used (Salvetat et al. 2006). Adopting D = 1.5 eV and7.5 × 10−8 g·cm−2 for the surface density (ρh), the data are well reproduced.

3.3 PAH temperature

The far-IR absorption band strengths are generally an order of magnitude smallerthan the mid-IR absorption band strengths. However in the common astronomicalcase of emission, relative band strengths can change significantly depending on thePAH’s internal vibrational energy (temperature).

In the framework of the thermal model, conservation of energy leads to thefollowing expression:

4π∑

i

σi

Tmax∫Ti

B(νi, T )[dT

dt

]−1

dT = Einternal (3.3)

where σi is the absorption cross-section in mode i; B(νi, T ), is Planck’s function atfrequency νi in mode i, at temperature T ; dT/dt is the cooling-rate and Einternal isthe PAH’s internal vibrational energy after excitation by a FUV photon. The sumis taken over all modes and the integral runs from the initial PAH temperature upto the PAH’s maximum attained temperature (Ti – Tmax) after absorption of theexciting photon.


Fig. 4. Comparison of the synthesized

absorption spectrum (10 – 1000 μm) of

coronene (bottom) with the emission spec-

trum produced with an internal vibrational

energy of 0.1 eV, both when approximating

the calculation of the heat capacity (mid-

dle) and exactly calculating the heat ca-

pacity (top; see also Boersma et al. submit-

ted). Bands have been given Lorentzian

profiles with a FWHM of 6 cm−1.

The maximum attained temperature is directly related to the energy of theabsorbed photon through the PAH heat capacity, which also effects its cooling-rate. The heat capacity can be calculated exactly by treating the PAH moleculeas a molecular system of isolated harmonic oscillators or approximated (see e.g.,Dwek et al. 1997).

Applying the thermal model to neutral coronene with an internal vibrationalenergy of 0.1 eV results in a maximum temperature of 189 Kelvin when calculat-ing the full heat capacity exactly and 201 Kelvin when using the approximation.Comparing both methods in Figure 4 shows the approximation is reasonably.

In absorption the coronene “drumhead” band intensity at 124 cm−1 is 2.9%of that of the strongest band at 863 cm−1, whereas at 201 K it has 60% of theintensity of the band at 863 cm−1. However, it should be kept in mind that theintrinsic strength of these bands is still weak.

3.4 Observing PAHs in the far-IR

Far-IR spectra of astronomical objects showing the mid-IR PAH emission featureswill provide a very different perspective on the PAH population than do the mid-IR bands, e.g., local conditions, chemical history of carbonaceous material, etc.Because the far-IR probes much colder material than the mid-IR, these data willexplore the large end of the astronomical PAH-size-distribution and thus expandthe scope of astronomical and astrochemical regimes in which PAHs can serve asprobes.

The far-IR bands promise to put a firm limit on the size of the emitting PAHs.However, the molecular structure of the PAH does not play a major part in de-termining the far-IR spectrum and, therefore, identifying specific molecules in theinterstellar PAH family likely requires other means. Mulas et al. (2006) point tothe rotational PQR band structure of the lowest PAH mode as such a means.Joblin et al., elsewhere in this volume, address an observational strategy based onthis band structure.


Calculations have shown that the lowest vibrational transitions of interstellarPAHs probably contain a few tenths of a percent of the absorbed FUV energy(Mulas et al. 2006). While this is only a tiny fraction, the decrease in backgrounddust continuum emission expected for warm PDRs could enhance the spectroscopiccontrast. One first-look observing strategy could be to study the emission fromtransition zones in harsh environments. Telescope sensitivity will then not be anissue and UV photolysis may have isomerised the PAHs to their most stable form,leaving only a very limited number of different species. Given their high stabilityand likely contribution to the mid-IR (Bauschlicher et al. 2008; Mattioda et al.2009), the presented coronene and pyrene “families” are particularly interesting inthis regard.

References



Boersma, C., Bauschlicher, C.W., Ricca, A., et al., ApJ, submitted

Dwek, E., Arendt, R.G., Fixsen, D.J., et al., 1997, ApJ, 475, 565

Malloci, G., Joblin, C., & Mulas, G., 2007, Chemical Phys., 332, 353

Mattioda, A.L., Hudgins, D.M., Boersma, C., et al., in preparation

Mattioda, A.L., Ricca, A., Tucker, J., Bauschlicher, C.W., & Allamandola, L.J., 2009,ApJ, 137, 4054

Meirovitch, L., 1997, Principles and techniques of vibrations (Prentice-Hall International)

Mulas, G., Malloci, G., Joblin, C., & Toublanc, D., 2006, A&A, 460, 93

Salvetat, J.P., Desarmot, G., Gauthier, C., & Poulin, P., 2006, Lecture Notes in Physics,Vol. 677, Mechanical Properties of Individual Nanotubes and Composites (SpringerBerlin/Heidelberg), 439


ANALYZING ASTRONOMICAL OBSERVATIONS WITH THENASA AMES PAH DATABASE

J. Cami1, 2

Abstract. We use the NASA Ames Polycyclic Aromatic Hydrocarbon(PAH) infrared spectroscopic database to model infared emission ofPAHs following absorption of a UV photon. We calculate emissionspectra resulting from the full cooling cascade for each species in thedatabase. Using a least squares approach, we can find out what PAHmixtures best reproduce a few typical astronomical observations rep-resenting the different classes of UIR spectra. We find that we canreproduce the observed UIR spectra in the wavelength range 6–14 μm,offering support for the hypothesis that the UIR bands are indeed dueto vibrational modes of PAHs and related molecular species. Spec-tral decompositions of our best fit models confirm and reinforce severalearlier results: (i) the 6.2 μm band requires a significant contributionof nitrogen-substituted PAHs (PANHs); (ii) the reported componentsand their variations in the 7.7 μm band are indicative of changes in thesize distribution of the contributing molecules; (iii) there is a signifi-cant contribution of anions to the 7.7 μm band; (iv) the 11.2 μm bandis due to large, neutral and pure PAHs; (v) the 11.0 μm band is due tolarge PAH cations.

1 Introduction

The PAH hypothesis that was introduced 25 years ago states that the so-calledUnidentified InfraRed (UIR) bands result from PAH molecules that are transientlyheated by absorption of a UV photon, and subsequently cool by emission of in-frared photons through vibrational relaxation (IR fluorescence). Although thishypothesis has now been generally accepted by the astronomical community, agood and detailed spectral match between a PAH mixture and astronomical ob-servations has not yet been produced. At the same time, it is not clear exactly

1 Department of Physics and Astronomy, University of Western Ontario, London, OntarioN6A 3K7, Canada2 SETI Institute - 189 Bernardo Ave., Suite 100 - Mountain View, CA 94043, USA



what the properties are of the astronomical PAH population. Such information isvital for astronomers if they are to use the widely observed UIR bands as diagnos-tic tools for probing the wide variety of astrophysical environments in which theUIR bands are observed.

Here, we present the first results of a program that uses the NASA Ames PAHinfrared spectroscopic database to analyze astronomical observations. We showthat a mixture of PAHs can indeed reproduce the observed UIR emission bandsand their spectral variations. Furthermore, we gain some insight into the natureof astronomical PAHs from spectral decompositions of the best fit models.

2 Modeling the PAH emission

2.1 PAH species

The NASA Ames PAH IR spectroscopic database (Bauschlicher et al. 2010; seealso Boersma et al. elsewhere in this volume) is currently the largest collectionof infrared properties of PAHs and some related molecules (e.g. the nitrogen-substituted PANHs, see Hudgins et al. 2005). The database contains theoreticallycalculated properties for a much larger set of molecules than is experimentallyavailable; here, we used the theoretical database exclusively.

Amongst others, the theoretical part of the database contains the calculatedfrequencies and intensities for all the fundamental mode vibrations for a collectionof 556 individual species. The database is highly biased toward smaller (≤30 Catoms) and pure PAHs that are neutral or singly ionized. We show in Section 4that reliable conclusions can still be inferred from such a biased database.

2.2 PAH physics

It is well established (see e.g. d’Hendecourt et al. 1989) that the PAH excitationand IR fluorescent cooling can be adequately described by using the thermal ap-proximation which states that the vibrational energy contained in a large, isolatedmolecule can be represented by a single parameter – the vibrational excitationtemperature. Absorption of a single UV photon will then briefly heat a moleculeto very high temperatures, after which it cools by infrared emission that scaleswith the Planck function at the vibrational temperature (see also Allamandolaet al. 1989). The total PAH emission following absorption of a single UV photonis then the time-integrated infrared emission as the molecule cools down.

The maximum temperature that can be attained by a PAH molecule after UVabsorption depends on the energy of the absorbed photon, and on the moleculecapability of distributing the energy over the vibrational states – the “vibrational”heat capacity. Since larger molecules have more vibrational modes, the averageenergy per vibrational mode is lower for a given photon energy and thus the largermolecules have typically lower temperatures after absorption of the same photon.As a consequence, they will emit much of their energy at longer wavelengths thanthe smaller PAHs.

J. Cami: Fitting the Astronomical PAHs 119

To take such effects properly into account, we used the data in the PAHdatabase to first calculate the heat capacities for each individual species. We alsocalculated the maximum temperature for each species upon absorption of a 10 eVphoton, and calculated the cooling rate for each species for temperatures between0 K and the maximum attained temperature (see e.g. Tielens 2005). This thenallows to calculate the full fluorescent infrared emission for each molecule in thedatabase by integrating the emission over the cooling curve. A check is performedto ensure energy conservation in the process.

2.3 PAH parameters

The PAH physics described above thus sets the relative intensities for all funda-mental modes for each molecular species individually, and the only free parameteris the energy of the absorbed UV photon. However, in order to compare thesemodel spectra to astronomical observations, we need to adopt a band profile shape.Most of the astronomical UIR bands are characterized by asymmetric band pro-files that are believed to originate from anharmonicities in the potential energy.This in turn results in hot band transitions that are slightly shifted in frequencycompared to the fundamental, and causes the overall profile to be asymmetric.

Since these effects reflect the details of the potential energy surface, they aregenerally different for each vibrational mode and for each molecule, and the result-ing band asymmetry is furthermore a function of the temperature. Unfortunately,the PAH database does not contain information on anharmonicity and integratinganharmonic effects is a very demanding task that is restricted so far to a few smallspecies (see Pirali et al. 2009; Basire et al. elsewhere in this volume). We thus de-cided to convolve each of the individual PAH bands with a symmetric Lorentzianprofile. We acknowledge that this will necessarily introduce some uncertainty inthe wings of the PAH bands; and we will therefore not discuss spectral features atthat level.

Finally, we only consider absorption by photons of one single energy – theaverage photon energy for the environment studied. The above procedures thusresult in a set of 556 individual spectra, representing the response of each speciesin the database to absorption by a photon with the same energy.

3 Fitting the astronomical spectra

3.1 Observations

Peeters et al. (2002) and van Diedenhoven et al. (2004) have shown that mostastronomical observations of the PAH bands can be divided in only a few classes(Class A, B and C; see also Peeters et al. elsewhere in this volume). Whereasobjects in class A have identical spectra, classes B and C represent a much largervariety in terms of peak positions of individual features and relative intensities ofthe PAH bands. For this work, we selected a representative object for class A (theHii region IRAS 23133+6050), but somewhat arbitrarily picked an object from


Fig. 1. (a) The three astronomical observations (solid lines) and the adopted dust con-

tinua (dashed lines). The Hii region IRAS 23133+6050 (top) is a Class A source; the Red

Rectangle (middle) a Class B source and the proto-planetary nebula IRAS 13416-6243

(bottom) a Class C object. (b) The astronomical PAH spectra (black) and the best fit

PAH model spectra (ligther curves). Residuals are shown as dashed lines.

class B (the Red Rectangle) and another one from class C (the proto-planetarynebula IRAS 13416-6243). If we can reproduce the observable UIR propertiesfor these three objects with PAHs, it stands to reason that we can equally wellreproduce most other observations.

All observations used here are ISO/SWS observations; we refer to Peeters et al.(2002) for details about the data reduction process. We do note here however thatwe chose to define a continuum that only represents the thermal emission of dust,and thus does not contain any spectral features to represent possible plateausunderneath the PAH features (see Fig. 1a). For what follows, we will always usethe continuum-subtracted spectra.

3.2 Procedures

We calculated individual PAH spectra for each of the species in the database asdescribed above, using an average photon energy of 10 eV. Then, we used a non-negative least squares algorithm to determine “abundances” (weights) for each ofthose PAHs such that the weighted sum of the PAH spectra provides the bestpossible fit to each of these astronomical observations.

One issue that needs to be addressed here is the intrinsic widths for each of thePAH bands in our models. Often, it is assumed that the intrinsic PAH bands arefairly broad (∼30 cm−1). However, the astronomical observations of the strongC-H out of plane bending mode at 11.2 μm show a band width (FWHM) of onlyabout 4 cm−1. Obviously, the constituent PAH spectra that make up this bandcannot be broader than this observed width. The same holds at least for the bandat 11.0 μm. In the 6–10 μm region on the other hand, the observed band profilesare broader (in wavenumber space), and it thus seems unrealistic to adopt a smallband width for the PAH profiles in this wavelength range. With some trial and

J. Cami: Fitting the Astronomical PAHs 121

Fig. 2. The spectral decomposition of our best fitting model as a function of molecular

size (left) and charge state (right). Astronomical observations are shown in black; overall

fits in red, and decompositions in green, blue or magenta (see legend).

error, we found that in this region, a width (FWHM) of 13 cm−1 provides betterglobal results. Thus, we convolved the individual PAH spectra with a Lorentzianprofile that has a FWHM of 13 cm−1 in the 6–10 μm region, and a width of 4 cm−1

in the 10–14 μm region. The resulting best fits are shown in Figure 1b.

4 Results & discussion: Fits & decomposition

Figure 1b clearly shows that a mixture of PAHs can indeed reproduce the observedinfrared emission for all three classes of astronomical PAH spectra. Typically,about 50 different molecular species are included in the models presented here.

Much more valuable than the fits themselves is the insight we can gain into thenature of the astrophysical PAH population by breaking down our best fit modelsinto spectral decompositions.

Figure 2 shows just two such examples. In the left panel, the green curverepresents the contribution to our total fit of the “large” PAHs, i.e. the species inour database with 30 or more carbon atoms; the blue curve represents the smallPAHs. This figure shows several interesting results. First, for both the Class Aand B sources, the strong 11.2 μm PAH band is (almost) exclusively due to largePAH species. This is consistent with the band width of 4 cm−1 that implies amean excitation temperature of about 280 K (Pech et al. 2002) and thus excludesa significant contribution of small PAHs that are too hot. Note that the NASAAmes PAH database is currently heavily biased toward smaller species, and giventhat bias, it is somewhat surprising to find that a spectral feature is exclusivelydue to large PAHs. Clearly, not a single one of the hundreds of small PAH speciesin the database is able to reproduce the 11.2 μm band without causing additionalspurious spectral features in this wavelength range. Also the PAH bands at 8.6


and 11.0 μm are due to large PAHs, and the same holds for the red componentof the 7.7 μm complex. Variations in the shape of the 7.7 μm complex are thenindicative of changes in the size distribution of the PAHs. Note also that the ClassC source contains what appears to be a random distribution of both large andsmall PAHs.

In the right panel, the best fit is decomposed according to the charge state ofthe contributing species: neutrals are shown in blue, cations in green, and anionsin magenta. Neutral PAHs produce the bulk of the emission in the 11.2 μm band,but cations are responsible for the 11.0 μm band. Combined with the left panel,we can thus conclude that the 11.2 μm band is due to large neutral PAHs, whilethe 11.0 μm band is due to large PAH cations. It is also interesting to see thatour best model includes a significant fraction of anions in the mixture. This isespecially clear in the 7.7 μm complex. Finally, note that for the class C source,most of the PAH emission arises from neutral molecules.

Many more spectral decompositions are possible, and yield interesting re-sults. One in particular is a decomposition according to the chemical compo-sition which shows that the 6.2 μm band requires a significant contribution ofnitrogen-substituted PAHs (called PANHs) – as was shown already by Hudginset al. (2005).

The work presented here illustrates the analytic and diagnostic power thatis contained in large spectral databases such as the NASA Ames PAH IR spec-troscopy database. The fact that some of our results from the spectral decompo-sition analysis go against the bias in the database (e.g. exclusively large PAHs inthe results while the database contains primarily small PAHs) indicates that ourresults are not driven by the database contents, but rather by the spectral proper-ties of the astronomical PAHs. At the same time, we are somewhat limited by therelatively small number of molecules in the database, and by the uncertainties inthe intrinsic band profiles. Nevertheless, the NASA Ames PAH database clearlycontains already a significant number of astrophysically relevant PAH species, andfuture research will undoubtedly increase the diagnostic value of exercises such asthe one presented here.

References



d’Hendecourt, L., Leger, A., Boissel, P., & Desert, F., 1989, IAUS, 135, 207

Hudgins, D.M., Bauschlicher, C.W. Jr., & Allamandola, L.J., 2005, ApJ, 632, 316


Peeters, E., Hony, S., van Kerckhoven, C., et al., 2002, A&A, 390, 1089


Tielens, A.G.G.M., 2005, The Physics and Chemistry of the Interstellar Medium(Cambridge University Press)

van Diedenhoven, B., Peeters, E., van Kerckhoven, C., et al., 2004, ApJ, 611, 928


SEARCH FOR FAR-IR PAH BANDS WITH HERSCHEL:MODELLING AND OBSERVATIONAL APPROACHES

C. Joblin1, 2, G. Mulas3, G. Malloci3 and E. Bergin & the HEXOSconsortium4

Abstract. Herschel opens the possibility to detect the low-frequency vi-brational bands of individual polycyclic aromatic hydrocarbon (PAH)molecules and therefore to progress in our understanding of the natureof these species and the properties of the environments from which theyemit. However, unless one individual molecule dominates the PAHfamily, this detection will not be straightforward and it is necessary tooptimise the observational search with an educated guess of the bandprofiles and intensities. Such educated guess can be obtained frommodels that include a detailed description of the molecular properties(anharmonicity, rotation...) in the modelling of the cooling cascade ofthe emitting species. First results are expected soon from the obser-vation of the Orion Bar as part of the HEXOS Herschel key program.

1 Introduction

Polycyclic aromatic hydrocarbons (PAHs) are commonly believed to be the car-riers of the aromatic infrared bands (AIBs) that dominate the mid–IR emissionbetween 3.3 and 15 μm in most galactic and extragalactic objects. However thePAH model suffers from the lack of identification of any individual species anddifferent approaches have been pursued over the years to address this question.Diffuse interstellar bands (DIBs), that fall essentially in the visible range (Herbig1995), are a way to identify individual PAH molecules by their low-lying electronictransitions, which are very characteristic. However, “fishing” for candidates has

1 Universite de Toulouse, UPS, CESR, 9 Av. colonel Roche, 31028 Toulouse Cedex 4, France2 CNRS, UMR 5187, 31028 Toulouse, France3 INAF – Osservatorio Astronomico di Cagliari – Astrochemistry Group, Strada 54, Loc.Poggio dei Pini, 09012 Capoterra (CA), Italy4 Department of Astronomy, University of Michigan, 500 Church Street, Ann Arbor,MI 48109, USA



not been very successful (see Snow & Destree and Cox and references in thesearticles, elsewhere in this volume). This approach is a lengthy task consideringthat each species has specific electronic transitions whose measurement for DIBidentification requires dedicated set-ups in the laboratory, to mimic the physicalconditions of low pressure and temperature in interstellar space. These experi-ments are very demanding (see for instance Tan & Salama 2006; Useli–Bacchittaet al. 2010; Pino et al., elsewhere in this volume) and the number of candidatesthat need to be investigated is very large. For a given number of carbon atoms,NC , there are different isomeric forms as well as a distribution of hydrogenationand ionisation states (cf. Bierbaum et al. and Montillaud et al., elsewhere in thisvolume).

Rotational spectroscopy is another technique that has been successful over theyears to identify circumstellar and interstellar molecules and radicals with the useof radio telescopes. Dipolar PAHs can display a pure rotational emission spectrumin the radio domain but transitions are expected to be weak; PAHs ought to berotating suprathermally in photodissociation regions (PDRs), yielding a very largerotational partition function (Rouan et al. 1992; Mulas 1998). Combined emissionsfrom different species will lead to forests of weak rotational lines that would mergeinto a quasi–continuum; PAHs are likely to contribute to the excess microwaveemission (Ysard et al. 2010). A better characterization of the nature of thecarriers of the anomalous emission would allow to constrain their spatial variationthat is required to retrieve the cosmic microwave background in the studies thatare currently led with the Planck mission.

Searches for the vibrational emission features in the far–IR range provide agood alternative to visible (DIB) and radio (rotational) identification methods andthat is becoming feasible with the Herschel mission. These features involve thewhole carbon skeleton of the molecule and are therefore more specific to the exactmolecular identity than the ones in the mid–IR, which instead probe functionalgroups. Experimental data in the far–IR range are still scarce and await for furtherprogress in the coming years (Zhang et al. 2010; Mattioda et al. 2009; Pirali et al.2006). On the other hand, more and more data are available from quantum-chemistry calculations using density functional theory (DFT; cf. Pauzat in thisvolume) and are collected in databases (Bauschlicher et al. 2010; Malloci et al.2007). This opens the possibility to model the far-IR emission of PAHs in PDRs.

2 Model of the (far–)IR emission spectrum of PAHs

The IR emission from vibrationally excited PAHs can be calculated from funda-mental equations (cf. Allamandola et al. 1989). The main step is to properly de-scribe the populations in all levels v of all modes i during the cooling process. Mostof the authors in the literature have used the simplifying assumption of thermaldistributions, in which the populations of vibrational levels are described by theBoltzmann equation, with the canonical formalism (cf. for instance Desert et al.1990; Pech et al. 2002; Draine & Li 2007). Our studies (Mulas 1998; Joblin et al.2002; Mulas et al. 2006a) are based on the microcanonical formalism in which the

C. Joblin et al.: Far-IR PAH Bands with Herschel 125

energy is the well–defined, conserved physical parameter. This more accuratelydescribes the statistical behaviour of a PAH molecule with a given internal energyU but comes at the price of having to explicitly take into account all vibrationalmodes. A Monte Carlo technique is used to properly solve the master equation andfollow the population of states during the cooling of the PAH down to the groundlevel. Other studies (cf. Ysard et al. 2010) used the exact-statistical method ofDraine & Li (2007). Comparison of results obtained respectively with the micro-canonical formalism and thermal approximation can be found in Joblin & Mulas(2009). Significant differences have been found for the band intensities (a factor oftwo or more for some of the bands). The emission spectrum of a large sample ofPAHs in different charge states (neutrals and cations) and in specific objects (suchas the Red Rectangle nebula) has been computed by Mulas et al. (2006a). TheMonte–Carlo code takes as an input molecular parameters that are available in thetheoretical on–line spectral database of PAHs (Malloci et al. 2007): the positionsand intensities of all vibrational modes and the photo–absorption cross–sectionsup to the vacuum–UV. This allows an accurate calculation of the band intensitiesbut the band profiles have to be assumed (cf. Fig. 1 and Mulas et al. 2006a).

The far–IR bands are emitted mainly at the end of the cooling cascade. Fur-thermore, at low excitation, it is necessary to consider that at some point theredistribution of energy between vibrational levels (the so-called internal vibra-tional redistribution, IVR) breaks down; the vibrational states decouple and theirquantum numbers cannot be described any more in statistical mechanics terms,but must instead be described explicitly in terms of specific vibrational transitionsbetween specific states. The population of states at decoupling is governed onlyby statistics and therefore does not depend on the Einstein coefficients associatedwith the far–IR modes. However the latter govern the relaxation rate and play akey role in the competition with the UV absorption rate. If the UV rate is slowcompared to the relaxation then the far–IR cooling can further proceed and veryweakly active or even inactive modes can relax in a manner analogous to the emis-sion of forbidden atomic lines in nebulae. In other cases, absorption of anotherUV photon triggers again mid–IR emission from the hot PAH, and this leads toreduced emission in the far–IR. In particular, the “slowest” far–IR modes can betotally suppressed (cf. case of IRAS 21282+5050 in Appendix C of Mulas et al.2006a).

3 Detailed calculation of the (far–) IR band profiles

A key aspect in the detection and identification of far–IR bands due to PAHsis related to the exact band profile. During the cooling cascade of the excitedPAH, most of the transitions happen between vibrationally excited states. Due tovibrational anharmonicity, the position and width of the emitted bands will dependon the involved states (cf. Oomens and Basire et al. in this volume). Many far–IRbands are due to out–of–plane vibrational modes, for which one can distinguishtwo main parts in the ro–vibrational band profile: the Q branch, in which finestructure is expected due to anharmonicity, and the P and R branches, that carry


Fig. 1. Calculated infrared emission spectrum from a sample of 40 PAHs ranging in size

from C10H8 to C48H20 and with neutral or cationic charge, compared with the estimated

dust continuum in the Red Rectangle nebula. Different panels correspond to different

values for σPR, the width of the P and R rotational branches, and σQ, the width of

the central Q branch (cf. Mulas et al. 2006a for more details). The operation ranges

of the relevant astronomical facilities are overlaid for reference. Herschel will open the

possibility to identify PAHs by their far-IR bands.

the rotational information. The far–IR bands being preferentially emitted near theend of the cooling cascades, the fine structure of these bands should not be washedout by coupling between states, as it happens for mid–IR bands. In particular, forout–of–plane modes, which have sharp Q branches, this enhances spectral contrast,making their detection possible in spite of the expected strong dust continuumbackground and spectral crowding (cf. (b) and (d) in Fig. 1).

The modelling of the band profile requires prior knowledge of several molecularparameters: the χ matrix of anharmonic parameters (including a proper treatmentof resonances) and the variations of the rotational constants as a function of thevibrational state. These parameters are not easy to access by experiment or theory;a recent study has been performed on naphthalene, C10H8 (Pirali et al. 2009),that illustrates the interplay between high-resolution spectroscopy and numericalsimulations of the photophysics of PAHs to access these parameters.


In order to compare with astronomical observations, unless one can use directlyexperimental results of UV-induced IR fluorescence taken under experimental con-ditions close to the ones expected in space (but such results are not available yet),some modelling of the emission process is necessary. As previously explained inSection 2, the two commonly used strategies are the simpler, faster, but less accu-rate thermal approximation, or the more complicated, but more accurate micro-canonical approach that we implemented in the form of Monte Carlo simulations(cf. Mulas et al. 2006a, 2006b). In our Monte Carlo model, the cooling cascadesof the PAHs were simulated assuming collisions to be negligible, and furthermoreneglecting photodissociation. While photodissociation is expected to be relevant inthe ISM, it still occurs on a relatively small fraction of the UV absorption events,resulting in a small correction to the estimated IR emission. Photoionisation is abit more frequent for neutrals and negative ions, and was taken into account. Alarge enough number of UV absorption events and relative IR cascades were sim-ulated, to accumulate sufficient statistics on the emission process. The populationof rotational modes (in the semirigid rotor approximation), and the statistics ofanharmonic shifts were recorded in the simulations, enabling us to produce fullrovibrational structures.

4 PAH far–IR spectroscopy with Herschel

From an observational point of view, the search for the far–IR bands of PAHs isnow becoming possible with the Herschel space observatory (Pilbratt et al. 2010)that has been launched in 2009. Herschel is equipped with three instruments thatcover the far-IR and sub-mm ranges from typically 50 to 600 μm. Detailed pre-sentations of the three instruments, the Photodetector Array Camera and Spec-trometer (PACS; Poglitsch et al. 2010), the Spectral and Photometric ImagingREceiver (SPIRE; Griffin et al. 2010) and the Herschel-Heterodyne Instrumentfor the Far-Infrared (HIFI; de Graauw et al. 2010) can be found elsewhere. Thebest strategy to detect the PAH far–IR bands is to search for the Q branches asso-ciated with the out–of–plane modes. The PACS and SPIRE instruments providethe best compromise between resolving power and sensitivity to evidence thesefeatures. Follow-up observations at very high resolution with the heterodyne spec-trometer HIFI will allow us to resolve the hot band structure of the Q branchesand may be also some structure in the P and R branches. These structures, ifdetected, carry detailed information on the emitting molecule as well as on thelocal physical conditions that prevail in the emitting environments. To retrievethis information it is necessary to model these features as described in Section 3.

As an example, we present here new calculations of band profiles performed onthe large molecule neutral circumovalene (C66H20). These calculations were mo-tivated by the Herschel observations of EXtra-Ordinary Sources (HEXOS; Berginet al. 2010) key program. In particular, observations are performed on the pro-totype bright PDR, the Orion Bar, following the strategy explained above. Forthe molecular parameters, realistic χ matrices were built based on the knowledgeof the smaller molecules C10H8 and C14H10 studied by Mulas et al. (2006b). A


Fig. 2. One example of the calculated structure of the low energy IR-active band of

circumovalene (C66H20), computed for one molecule embedded in the Orion Bar radiation

field assumed to be G0 = 5×104 in Habing units. Full PQR structure (left) and zoom in

on the Q branch (right). Green (upper) spectrum is at full resolution and illustrates the

complex structure that could be revealed by the Herschel HIFI heterodyne spectrometer.

Blue (lower) spectrum has been convolved at the highest resolution of the Herschel SPIRE

instrument.

similar educated guess was performed for the variations of the rotational constants.The radiation field of the Orion Bar was calculated using the Meudon PDR code(Le Petit et al. 2006) with a stellar spectrum from the Kurucz library with atemperature of 40 000 K (Kurucz 1991), a distance of 0.25 pc and a gas densityof nH = 5 × 104 cm−3. The calculated radiation field at the cloud surface has anintegrated UV intensity of G0 = 5×104 in Habing units (Habing 1968). The cool-ing cascades of C66H20 in the Orion Bar field were then calculated and the bandprofiles obtained as described in Section 3. The calculated profiles were convolvedat the spectral resolution of the instruments onboard Herschel (cf. Fig. 2).

5 Conclusions and perspectives

The search for the far–IR bands of PAHs is expected to be a difficult subject sincethe bands are predicted to be weak and our ability to observe them will dependon how diverse is the mixture of interstellar PAHs. Herschel offers the possibilityto investigate this question thanks to the characteristics and the spectral rangecovered by its instruments. SOFIA and ALMA that are now entering the scenecould also bring new insights into this topics. The next far-IR/sub-mm spacemission will then be SPICA (2018). But none of these new missions will allow tocover the full Herschel spectral range and we should therefore try hard to search


for these PAH bands with Herschel. As we have illustrated, a proper modelling ofthe cooling cascade including molecular properties (anharmonicity, rotation...) ismandatory for band prediction and analysis. In this regard, there is still a lot ofwork to do in the laboratory and this not an easy task for either the experimentor the theoretical calculations. From the experimental point of view, there isa need for far–IR spectroscopy of large PAH molecules, but at T < 300 K toresolve (some) of the anharmonic structure. The calculations of the matrix ofanharmonic parameters with DFT are very heavy and “tricky”. Futhermore, forhigh temperatures, alternatives have to be found such as molecular dynamics withan approximated DFT scheme (Porezag et al. 1995; see also Pauzat and Basireet al., elsewhere in this volume). A lot of effort still needs to be dedicated to thissubject both from the observational and laboratory points of view.

References


Bauschlicher, C.W., Boersma, C., Ricca, A., Mattioda, A.L., et al., 2010, ApJS, 189, 341

Bergin, E.A., Phillips, T.G., Comito, C., et al., 2010, A&A, 521, L20

Desert, F.-X., Boulanger, F., & Puget, J.L., 1990, A&A, 237, 215

Draine, B.T., & Li, A., 2001, ApJ, 551, 807

Draine, B.T., & Li, A., 2007, ApJ, 657, 810

de Graauw, T., Helmich, F.P., Phillips, T.G., et al., 2010, A&A, 518, L4

Griffin, M.J., Abergel, A., Abreu, A., et al., 2010, A&A, 518, L3

Habing, H.J., 1968, BAIN, 19, 421

Herbig, G.H., 1995, ARA&A, 33, 19

Joblin, C., Boissel, P., Leger, A., D’Hendecourt, L., & Defourneau, D., 1995, A&A, 299,835

Joblin, C., Toublanc, D., Boissel, P., & Tielens, A.G.G.M., 2002, Mol. Phys., 100, 3595

Joblin, C., & Mulas, G., 2009, in: Interstellar dust properties: from fundamental studiesto astronomical models, ed. F. Boulanger, C. Joblin, A. Jones, & S. Madden, EASPub. Series, 35, 133

Kurucz, R.L., 1991, BAAS, 23, 1047

Le Petit, F., Nehme, C., Le Bourlot, J., & Roueff, E., 2006, ApJS, 164, 506


Mattioda, A.L., Ricca, A., Tucker, J., Bauschlicher, C.W., & Allamandola, L.J., 2009,AJ, 137, 4054

Mulas, G., 1998, A&A, 338, 243

Mulas, G., Malloci, G., Joblin, C., & Toublanc, D., 2006a, A&A, 460, 93

Mulas, G., Malloci, G., Joblin, C., & Toublanc, D., 2006b, A&A, 456, 161


Pilbratt, G.L., Riedinger, J.R., Passvogel, T., et al., 2010, A&A, 518, L1

Pirali, O., Van-Oanh, N.-T., Parneix, P., Vervloet, M., & Brechignac, P., 2006, Phys.Chem. Chem. Phys., 8, 3707



Poglitsch, A., Waelkens, C., Geis, N., et al., 2010, A&A, 518, L2

Porezag, D., Frauenheim, T., Kohler, T., Seifert, G., & Kaschner, R., 1995, Phys. Rev.B, 51, 12947

Rouan, D., Leger, A., Omont, A., & Giard, M., 1992, A&A, 253, 498

Salama, F., Galazutdinov, G.A., Kre�lowski, J., Allamandola, L.J., & Musaev, F.A., 1999,ApJ, 526, 265


Zhang, J., Han, F., Pei, L., Kong, W., & Li, A., 2010, ApJ, 715, 485

PAHs and Star Formation

in the Near and Far Universe


POLYCYCLIC AROMATIC HYDROCARBONS AS STARFORMATION RATE INDICATORS

D. Calzetti1

Abstract. As images and spectra from ISO and Spitzer have providedincreasingly higher–fidelity representations of the mid–infrared (MIR)and Polycyclic Aromatic Hydrocarbon (PAH) emission from galaxiesand galactic and extra–galactic regions, more systematic efforts havebeen devoted to establishing whether the emission in this wavelengthregion can be used as a reliable star formation rate indicator. Thishas also been in response to the extensive surveys of distant galaxiesthat have accumulated during the cold phase of the Spitzer Space Tele-scope. Results so far have been somewhat contradictory, reflecting thecomplex nature of the PAHs and of the mid–infrared–emitting dust ingeneral. The two main problems faced when attempting to define astar formation rate indicator based on the mid–infrared emission fromgalaxies and star–forming regions are: (1) the strong dependence ofthe PAH emission on metallicity; (2) the heating of the PAH dust byevolved stellar populations unrelated to the current star formation. Ireview the status of the field, with a specific focus on these two prob-lems, and will try to quantify the impact of each on calibrations of themid–infrared emission as a star formation rate indicator.

1 Introduction

First enabled by the data of the Infrared Space Observatory (ISO, Kessler et al.1996) and then expanded by the images and spectra of the Spitzer Space Telescope(Spitzer, Werner et al. 2004), the mid–infrared, ∼3–40 μm, wavelength range hasbeen at the center of many investigations seeking to define handy and easy–to–useStar Formation Rate (SFR) indicators, especially in the context of extragalacticresearch.

In the Spitzer (cold–phase) era, the MIR emission from galaxies had experi-enced renewed interest particularly in the high–redshift galaxy populations com-munity (e.g., Daddi et al. 2005,2007; Papovich et al. 2006,2007; Yan et al. 2007;

1 Dept. of Astronomy, University of Massachusetts, Amherst, MA 01003, USA,e-mail: [email protected]



Reddy et al. 2010). The Spitzer Multiband Imaging Photometer (MIPS, Riekeet al. 2004) and the Spitzer InfraRed Spectrograph (IRS, Houck et al. 2004) hadproven particularly sensitive at detecting the MIR dust emission from galaxies inthe redshift range z≈ 1–3. Observations performed with the MIPS 24 μm and70 μm bands yield the restframe ∼8 μm and 23 μm emission of a z=2 galaxy. Anobvious question to ask is whether, and to what degree, the MIR emission tracesthe recent star formation in galaxies.

The MIR region hosts the rich spectrum of the Polycyclic AromaticHydrocarbon (PAH) emission features in the wavelength range 3–20 μm (Leger& Puget 1984), also previously known as Unidentified Infrared Bands or AromaticFeatures in Emission (e.g., Sellgren et al. 1983; Sellgren 1984; Puget et al. 1985),which are the subject of the present Conference. Thus, a corollary to the ques-tion in the previous paragraph is how well the PAH emission traces recent starformation in galaxies.

In this Review, I concentrate on results obtained from the investigation ofnearby (closer than about 30–50 Mpc) galaxies, where the MIR emission can betypically resolved to sub–kpc scale at the Spitzer resolution in the 3–8 μm range:the 2′′ FWHM corresponds to ∼0.5 kpc for a galaxy at 50 Mpc distance. In thisregime, separating the contributing heating effects of various stellar populationsbecome easier than in unresolved galaxies. Throughout this paper, I will referto “8 μm emission” when indicating the stellar–continuum subtracted emissionas detected by Spitzer in the IRAC 8 μm band (Fazio et al. 2004) or similar–wavelength bands on the ISO satellite. With the terminology “PAH emission” Iwill refer to the emission features (after subtraction of both the stellar and dustcontinuum) at/around 8 μm.

An important assumption when using the infrared emission as a SFR indica-tor is that galaxies must contain sufficient dust that a significant fraction of theUV–optical light from recently formed stars is absorbed and re–emitted at longerwavelengths by the dust itself. For the “typical” galaxy in the Universe, abouthalf of its energy budget is re–processed by dust in the infrared (e.g., Hauser &Dwek 2001; Dole et al. 2006), but a very large scatter on this “mean” value existsfrom galaxy to galaxy. There tends to be a relation between star formation activ-ity and dust opacity in galaxies, in the sense that more active galaxies also tendto be more opaque (e.g., Wang & Heckman 1996; Heckman et al. 1998; Calzetti2001; Sullivan et al. 2001). In general, infrared SFR calibrations will need to takeinto account that a fraction of the star formation will emerge from the galaxyunprocessed by dust, and this fraction will depend on a number of factors (metalcontent and star formation activity being two of those, and dust geometry beinga third, harder–to–quantify, factor).

2 PAH emission as a SFR indicator

The 8 μm emission represents about 5%–20% of the total infrared (TIR) emis-sion from galaxies; of this, more than half, and typically 70%, can be attributedto PAH emission in metal–rich galaxies (Smith et al. 2007; Dale et al. 2009;

D. Calzetti: PAHs as SFR Indicators 135

Marble et al. 2010). In this context, “metal–rich” refers to oxygen abundanceslarger than 12+ log(O/H)∼8.1–8.2 (with the Sun’s oxygen abundance being ∼8.7,Asplund et al. 2009). Although small, the fraction of TIR energy contained inthe 7–8 μm wavelength region is not entirely negligible, and much effort has beendevoted to investigating whether the emission in this region could be used as areliable SFR indicator.

Systematic efforts began with ISO (e.g., Roussel et al. 2001; Boselli et al.2004; Forster–Schreiber et al. 2004; Peeters et al. 2004), and continued withSpitzer, which enabled extending the analysis to fainter systems and to higherspatial detail (e.g., Wu et al. 2005; Calzetti et al. 2005, 2007; Alonso–Herreroet al 2006; Zhu et al. 2008; Kennicutt et al. 2009; Salim et al. 2009; Lawtonet al. 2010). In a typical approach, the 8 μm (or similar wavelengths) luminosityor luminosity/area is plotted as a function of other known SFR indicators, toestablish whether a correlation exists; an example is shown in Figure 1. In onecase, the 8 μm emission has also been combined with tracers of the unattenuatedstar formation (e.g., Hα line emission) to recover an “unbiased” SFR indicator(Kennicutt et al. 2009).

Most of these analyses recover a linear (in log–log space) relation with slope ofunity, or slightly less than 1, between the 8 μm emission and the reference SFRindicator. A roughly linear trend with unity slope would in general indicate thatthe 8 μm luminosity is an excellent SFR tracer; however, as we will see in thenext two sections, the PAH emission (and, therefore, the 8 μm emission, of whichthe PAH emission represents typically more than half of the total) shows a strongdependence on the metal content of the region or galaxy, and a less strong, butpossibly significant, dependence on the nature of the heating stellar population.

3 The dependence of PAHs on metallicity

Already known for more than a decade (e.g., Madden 2000, and references therein),the dependence of the intensity of the PAH emission on the galaxy/environmentmetallicity has been quantified in detail by Spitzer data (e.g., Fig. 1). This de-pendency is on top of the general decrease of the TIR intensity for decreasingmetallicity, i.e. the decrease in overall dust content. Analyses of galaxy samplesand of galaxy radial profiles covering a range of metallicity show an additionalorder–of–magnitude decrease in the 8 μm–to–TIR luminosity for a factor ∼10 de-crease in metallicity, with a transition at 12+log(O/H)≈8.1 (Boselli et al. 2004;Madden et al. 2006; Engelbracht et al. 2005; Hogg et al. 2005; Galliano et al.2005,2008; Rosenberg et al. 2006; Wu et al. 2006; Draine et al. 2007; Engelbrachtet al. 2008; Gordon et al. 2008; Munoz–Mateos et al. 2009; Marble et al. 2010).

MIR spectroscopy with Spitzer has established that the observed low abun-dance of PAHs in low metallicity systems is not due to the molecules being morehighly ionized and/or dehydrogenated than in higher metallicity galaxies (Smithet al. 2007), although there is controversy on whether they could be characterizedby smaller or larger sizes (Hunt et al. 2010; Sandstrom et al. 2010; see, also,


Fig. 1. The surface luminosity density at 8 μm (stellar–continuum subtracted) as a

function of the surface luminosity density of the extinction–corrected hydrogen recombi-

nation emission line Pα (1.8756 μm), for HII knots (red, green, and blue symbols) and

low–metallicity starburst galaxies (black symbols). The HII-knots are typically ≈0.5 kpc–

size line–emitting regions from 33 nearby galaxies; there are a total of 220 regions in the

plot. The starburst galaxies are from the sample of Engelbracht et al. (2005). The

red symbols show regions hosted in galaxies with oxygen abundance 12+log(O/H)>8.35.

Regions of lower metal abundance are divided between those with oxygen abundance

8.00<12+log(O/H)≤8.35 (green symbols) and those with 12+log(O/H)≤8.00 (blue sym-

bols). The continuous line shows the best linear fit (in the log–log plane) through the high

metallicity data points (red symbols). The dash line shows the best linear fit with unity

slope. A common result is for the best fit to yield a slope lower than unity. The dot-dash

curve shows the locus of a model for which regions/galaxies suffer from decreasing dust

attenuation (and, consequently, decreasing dust emission) as their total luminosity de-

creases, a commonly observed trend in galaxies and star–forming regions (see references

in the last paragraph of the Introduction). Even when taking this effect into account,

low metallicity regions and starburst galaxies display a depressed 8 μm emission, which

has been shown to be due to deficiency of PAH emission. From Calzetti et al. (2007).

Hunt et al. elsewhere in this volume and Sandstrom et al. elsewhere in thisvolume).

The nature of the correlation between the strength of the PAH emission featuresand the region’s metallicity is still ground for debate. The two main scenarios


that can account for the observed trend are: the “nature” scenario, according towhich lower metallicity systems have a delayed PAH formation, and the “nurture”scenario, according to which PAHs are processed and/or destroyed in the harderradiation fields of low metallicity galaxies.

Galliano et al. (2008, see, also, Galliano elsewhere in this volume) suggestthat the PAH strength–metallicity correlation may be due to the delayed for-mation of the PAHs themselves, which are thought to form in the envelopes ofcarbon–rich AGB stars. In a young system, the first dust will emerge from su-pernovae (timescale<10 Myr), while AGB–produced dust will emerge at a laterstage (timescale≈1 Gyr). This scenario may be difficult to reconcile with the factthat low–metallicity systems in the local Universe contain stellar populations thatare typically older than 2 Gyr (Tosi 2009 and references therein), and have me-dian birth parameters that are not drastically different from those of more metalrich galaxies (Lee et al. 2007). However, low–mass systems are also thought tolose most of their metals during the sporadic events of star formation that char-acterize their typical star formation history (e.g., Romano et al. 2006). Clearly,self–consistent models for the metal enrichment of dwarf, low–metallicity galaxiesthat also include their extended star formation histories are needed in order toaccept or refute the “nature” scenario.

Alternately, processing of PAHs by hard radiation fields has been proposed asa mechanism for the observed correlation between PAH luminosity and metallicity(Madden et al. 2006; Wu et al. 2006; Bendo et al. 2006; Smith et al. 2007;Gordon et al. 2008; Engelbracht et al. 2008), since low–metallicity environmentsare generally characterized by harder radiation fields than high–metallicity ones(e.g. Hunt et al. 2010). Additionally, efficient destruction of PAHs by supernovashocks may also contribute to the observed deficiency (O’Halloran et al. 2006).These suggestions agree with the observation that PAHs are present in the PDRssurrounding HII regions, but are absent (likely destroyed) within the HII regions(Cesarsky et al. 1996; Helou et al. 2004; Bendo et al. 2006; Relano & Kennicutt2009). A recent study of the SMC finds a high fraction of PAHs within molecularclouds (Sandstrom et al. 2010), thus complicating the interpretation of the forma-tion/destruction mechanisms for these molecules. Both production and processingmay ultimately be driving the observed trend (Wu et al. 2006; Engelbracht et al.2008; Marble et al. 2010).

4 The heating populations of PAHs

Outside of HII regions and other strongly ionizing environments, PAHs tend tobe ubiquitous. These large molecules are transiently heated by single UV andoptical photons, and, therefore, they can be heated by the radiation from the mixof stellar populations that contribute to the general interstellar radiation field,unrelated to the current star formation in a galaxy (Haas et al. 2002; Boselli et al.2004; Peeters et al. 2004; Mattioda et al. 2005; Calzetti et al. 2007; Draine & Li2007; Bendo et al. 2008).


Bendo et al. (2008) show that the 8 μm emission from galaxies is better cor-related with the 160 μm dust emission than with the 24 μm emission, on spatialscales ≈2 kpc. Haas Klaas & Bianchi (2002) also find that the ISO 7.7 μm emis-sion is correlated with the 850 μm emission from galaxies. Boselli et al. (2004)show that the stellar populations responsible for the heating of the MIR–emittingdust are more similar to those responsible for heating the large grains emittingin the far–infrared, rather than to those responsible for the ionized gas emissionin galaxies. This result is similar to that obtained by Peeters et al. (2004) fora combined Galactic and extragalactic sample of PAH–emitting sources. All theevidence points to a close relation between the MIR/PAH emission and the colddust heated by the general (non–star–forming) stellar population.

The presence of heating by evolved stellar populations thus produces a ill–quantified contribution to the total MIR luminosity of a galaxy, affecting thecalibration of any SFR indicator using that wavelength range (e.g., Alonso–Herreroet al. 2006). We still do not have a clear understanding of dependencies on galaxymorphology, stellar population mix, star formation rate, star formation intensity,etc.

A recent analysis of the nearby galaxy NGC 0628, an almost face–on SAc at adistance of about 7.3 Mpc, shows that its 8 μm emission contains about 20%–30%contribution from a diffuse component unrelated to sites of current star formation(Crocker et al. 2010). The 20%–30% range reflects uncertainties in the adoptedextinction and [NII] contamination corrections for the Hα emission, used here totrace sites of recent star formation. We can also obtain a rough idea of the amountof diffuse emission in the MIR bands by converting the 8 μm mean luminositydensity, ∼1.2 × 107 L� Mpc−3, within the local 10 Mpc (including emission fromboth PAHs and dust continuum, see Marble et al. 2010) to a volume SFR density,using metallicity–dependent linear calibrations to SFR(8 μm) from the data ofCalzetti et al. (2007). The result, ρSFR(8 μm)∼0.019 M� yr−1 Mpc−3, is roughly30%–60% higher than the commonly accepted values for the SFR density in thelocal Volume (see references in Hopkins & Beacom 2006).

While the above suggests that the contribution to the 8 μm emission from dustheated by the diffuse stellar populations is not large (less than a factor ≈1.5–2),when galaxies are considered as a whole or as populations, we should recall thatwe don’t have a handle on galaxy–to–galaxy variations. Even worse, the diffusepopulation heating could become a prevalent contribution to the 8 μm emissionin some galactic regions, thus potentially affecting any investigation of spatiallyresolved features within galaxies.

5 Summary and conclusions

Over the past decade, the accumulation of both spectroscopy and high–angular–resolution imaging of the MIR emission from galaxies has paved the road for anaccurate investigation of the MIR/PAH luminosity as a SFR tracer. Data haveshown that there is generally a linear (with slope of 1 or slightly less than 1 in a log–log plane) correlation between the stellar–continuum–subtracted MIR luminosity


and the SFR, for high metallicity systems, i.e., for galaxies and regions that areabout 1/5–1/3 solar or higher in oxygen abundance. While a linear correlationwould generally indicate that the MIR emission can be considered a reliable SFRindicator, there are at least two caveats to keep in mind: (1) the PAH emissionhas a strong dependence on metallicity; (2) the MIR–emitting dust can be heatedby evolved stellar populations unrelated to the current star formation.

The PAH dependence on metallicity is well established and quantified, andshows an order–of–magnitude deficiency in PAH/TIR emission for a decrease inoxygen abundance by about a factor 10. The nature of this dependency is stilldebated, and could be due to delayed production of the PAHs in low metallicitysystems or to processing/destruction mechanisms in the harder radiation fields oflow metallicity galaxies, or a combination of both.

The heating of PAHs by evolved stellar populations is also well established,but less well quantified. It is likely to have a smaller impact on any SFR(PAH)or SFR(MIR) calibration than the metallicity dependence, probably at the levelof less than a factor 2. However, potential variations as a function of galaxymorphology, stellar population mix, star formation rate, star formation intensity,etc., and the impact of the diffuse population heating as a function of locationfor sub–galactic scale analyses have not been quantified yet. This is a virtuallyunexplored realm with a possibility for major implications as many investigationsmove from the “global populations” approach to the “sub–kpc regions” approachin the study of star formation in galaxies.

References

Alonso–Herrero, A., Colina, L., Packam, C., et al., 2006, ApJ, 652, L83

Aplund, M., Grevesse, N., Sauval, A.J., & Scott, P., 2009, ARA&A, 47, 481

Bendo, G.J., Dale, D.A., Draine, B.T., et al., 2006, ApJ, 652, 283

Bendo, G.J., Draine, B.T., Engelbracht, C.W., et al., 2008, MNRAS, 389, 629

Boselli, A., Lequeux, J., & Gavazzi, G., 2004, A&A, 428, 409

Calzetti, D., 2001, PASP, 113, 1449

Calzetti, D., Kennicutt, R.C., Bianchi, L., et al., 2005, ApJ, 633, 871

Calzetti, D., Kennicutt, R.C., Engelbracht, C.W., et al., 2007, ApJ, 666, 870

Cesarsky, D., Lequeux, J., Abergel, A., et al., 1996, A&A, 315, L309

Crocker, A.F., Calzetti, D., et al., 2010, in preparation

Daddi, E., Dickinson, M., Chary, R., et al., 2005, ApJ, 631, L13

Daddi, E., Dickinson, M., Morrison, G., et al., 2007, ApJ, 670, 156

Dale, D.A., Smith, J.T.D., Schlawin, E.A., et al., 2009, ApJ, 693, 1821

Dole, H., Lagache, G., Puget, J.-L., et al., 2006, A&A, 451, 417

Draine, B.T., & Li, A., 2007, ApJ, 657, 810


Engelbracht, C.W., Gordon, K.D., Rieke, G.H., et al., 2005, ApJ, 628, 29

Engelbracht, C.W., Rieke, G.H., Gordon, K.D., et al., 2008, ApJ, 685, 678

Fazio, G.G., Hora, J.L., Allen, L.E., et al., 2004, ApJS, 154, 10


Forster Schreiber, N.M., Roussel, H., Sauvage, M., & Charmandaris, V., 2004, A&A,419, 501

Galliano, F., Madden, S.C., Jones, A.P., Wilson, C.D., & Bernard, J.-P., 2005, A&A,434, 867



Haas, M., Klaas, U., & Bianchi, S., 2002, A&A, 385, L23

Hauser, M.G., & Dwek, E., 2001, ARA&A, 39, 249

Heckman, T.M., Robert, C., Leitherer, C., Garnett, D.R., & van der Rydt, F., 1998,ApJ, 503, 646

Helou, G., Roussel, H., Appleton, P., et al., 2004, ApJS, 154, 253

Hogg, D.W., Tremonti, C.A., Blanton, M.R., et al., 2005, ApJ, 624, 162

Hopkins, A.M., & Beacom, J.F., 2006, ApJ, 651, 142

Houck, J.R., Roellig, T.L., van Cleve, J., et al., 2004, ApJS, 154, 18

Hunt, L.K., Thuan, T.X., Izotov, Y.I., & Sauvage, M., 2010, ApJ, 712, 164

Kennicutt, R.C., Hao, C.-N., Calzetti, D., et al., 2009, ApJ, 703, 1672

Kessler, M.F., Steinz, J.A., Anderegg, M.E., et al., 1996, A&A, 315, L27

Lawton, B., Gordon, K.D., Babler, B., et al., 2010, ApJ, 716, 453

Lee, Janice, C., Kennicutt, R.C., Funes, J.G., Sakai, S., & Akiyama, S., 2007, ApJ, 671,L113


Madden, S.C., 2000, in Massive Stellar Clusters, ed. A Lancon, & C. Boily, ASP Conf.Ser., 211, 297


Marble, A.R., Engelbracht, C.W., van Zee, L., et al., 2010, ApJ, 715, 506

Mattioda, A.L., Allamandola, L.J., & Hudgins, D.M., 2005, ApJ, 629, 1183

Munoz–Mateos, J.C., Gil de Paz, A., Boissier, S., et al., 2009, ApJ, 701, 1965


Papovich, C., Moustakas, L.A., Dickinson, M., et al., 2006, ApJ, 640, 92

Papovich, C., Rudnick, G., Le Floc’h, E., et al., 2007, ApJ, 668, 45

Peeters, E., Spoon, H.W.W., & Tielens, A.G.G.M., 2004, ApJ, 613, 986

Puget, J.L., Leger, A., & Boulanger, F., 1985, A&A, 142, L19

Reddy, N.A., Erb, D.K., Pettini, M., Steidel, C.C., & Shapley, A.E., 2010, ApJ, 712,1070

Relano, M., & Kennicutt, R.C., 2009, ApJ, 699, 1125

Rieke, G.H., Young, E.T., Engelbracht, C.W., et al., 2004, ApJS, 154, 25

Romano, D., Tosi, M., & Matteucci, F., 2006, MNRAS, 365, 759

Rosenberg, J.L., Ashby, M.L.N., Salzer, J.J., & Huang, J.-S., 2006, ApJ, 636, 742

Roussel, H., Sauvage, M., Vigroux, L., & Bosma, A., 2001, A&A, 372, 427

Salim, S., Dickinson, M., Rich, R.M., et al., 2009, ApJ, 700, 161

Sandstrom, K.M., Bolatto, A.D., Draine, B.T., Bot, C., & Staminirovic, S., 2010, ApJ,715, 701

Sellgren, K., Werner, M.W., & Dinerstein, H.L., 1983, ApJ, 271, L13


Sellgren, K., 1984, ApJ, 277, 623


Sullivan, M., Mobasher, B., Chan, B., et al., 2001, ApJ, 558, 72

Tosi, M., 2009, A&A, 500, 157

Wang, B., & Heckman, T.M., 1996, ApJ, 457, 645

Werner, M.W., Roellig, T.L., Low, F.J., et al., 2004, ApJS, 154, 1

Wu, H., Cao, C., Hao, C.-N., et al., 2005, ApJ, 632, L79

Wu, Y., Charmandaris, V., Hao, L., et al., 2006, ApJ, 639, 157

Yan, L., Sajina, A., Fadda, D., et al., 2007, ApJ, 658, 778

Zhu, Y.-N., Wu, H., Cao, C., & Li, H.-N., 2008, ApJ, 686, 155


PAHS AND THE ISM IN METAL-POOR STARBURSTS

L.K. Hunt1, Y.I. Izotov2, M. Sauvage3 and T.X. Thuan4

Abstract. We characterize PAH populations in 22 metal-poor blue com-pact dwarf galaxies (BCDs), 16 of which have an oxygen abundance12+log(O/H) <∼ 8. This is the largest sample ever studied at suchlow metallicities. The relative PAH intensities of the 6.2, 7.7, 8.6 and11.3 μm features in these BCDs suggest a deficit of small PAH carriers,or alternatively, an excess of large ones at these low abundances.

1 Introduction

PAH emission dominates the mid-infrared spectra of star-forming galaxies (Brandlet al. 2006; Smith et al. 2007), and contributes importantly to their IR energybudget. In most star-forming galaxies, PAH emission has a surprisingly narrowrange of properties, well characterized by a “standard set” of features (Smith et al.2007). However, first ISO and later Spitzer have shown that metal-poor star-forming galaxies are deficient in PAH emission (Madden et al. 2006; Wu et al.2006; Engelbracht et al. 2008), similar to the observed deficit in CO emission (e.g.,Taylor et al. 1998). Neither PAH nor CO emission is generally detected below12+log(O/H)∼8, consistent with the idea that PAHs and CO form in similarconditions within molecular clouds (Sandstrom et al. 2010, see Sandstrom et al.elsewhere in this volume).

Although it has generally been concluded that the reason for this PAH de-ficiency is low metallicity, it is not yet clear whether it is the metallicity di-rectly, or rather the indirect effects of a low metallicity environment. Previouswork with ISO (e.g., Galliano et al. 2005; Madden et al. 2006) examined thisquestion, but the metallicity of the galaxies studied by ISO did not go below12+log(O/H)∼ 8.2, making it difficult to probe the physical conditions wherePAH emission is effectively suppressed. Spitzer/IRS added more observations of

1 INAF-Osservatorio di Arcetri, Firenze, Italy2 Main Astronomical Observatory, Kiev, Ukraine3 CEA, Saclay, France4 University of Virginia, Charlottesville, USA



the PAHs in low-metallicity dwarf galaxies, but up to now only five objects with12+log(O/H) <∼ 8 (Wu et al. 2006) were available for study.

Here, we present observations of PAH features in a new sample of 22 metal-poor blue compact dwarf galaxies (BCDs), 16 of which have 12+log(O/H) <∼ 8;this more than triples the number of such dwarfs with PAH spectra. We examinein detail the PAH emission in our sample, and show for the first time that somePAHs can survive even in extremely metal-poor environments.

2 Observations and analysis

In the context of our GO Spitzer program (PID 3139), we obtained IRS spectrain the low- and high-resolution modules (SL, SH, LH: Houck et al. 2004). Thesample definition and data reduction are described in detail by Hunt et al. (2010).In order to directly compare the results for our low-metallicity sample with thoseof more metal-rich galaxies, we analyzed our IRS spectra with PAHFIT (Smithet al. 2007), an IDL procedure which was developed for and applied to the SpitzerNearby Galaxy Survey (SINGS: Kennicutt et al. 2003). PAHFIT is particularlysuited for separating emission lines and PAH features (e.g., the PAH blend at12.6–12.7μm and the [Ne ii] line at 12.8μm), as well as measuring faint PAHfeatures superimposed on a strong continuum. PAHFIT overcomes this problemby fitting simultaneously the spectral features, the underlying continuum, and theextinction. Drude profiles are fitted to the PAH features, and Gaussian profiles tothe molecular hydrogen and fine-structure lines.

3 Results

The 7.7μm blend is the most common PAH feature in our sample, detected in 15of 22 objects. The 7.7μm feature is also the strongest one, comprising roughly 49%of the total PAH power. The remaining PAHs are significantly weaker, and lessfrequently detected: 13 BCDs show the 11.2–11.3μm PAH, 9 the 8.6μm feature,and 7 the 6.2μm PAH. No 17μm PAH features were detected. The 6.2, 7.7, 8.6,11.3, and 12.6μm features by themselves constitute ∼72% of total PAH power inBCDs; the SINGS sample has about ∼85% of total PAH power in these bands(Smith et al. 2007). The remainder of the emission is in the weaker features.

3.1 Fractional power of strongest features

The fractional power of the four strongest features relative to the total PAH lu-minosity is illustrated in Figure 1. The horizontal lines in each panel correspondto the SINGS medians (dashed) and the BCD means (dotted). The BCD meansare calculated taking into account all galaxies with 7.7μm detections; thus theyare a sort of weighted average which considers frequency of detection togetherwith intensity. This is why the mean for the 6.2μm PAH lies below most of theBCD data points: that feature was detected only in 7 galaxies, while the otherfeatures in Figure 1 were detected with a frequency more similar to the 7.7μmPAH. The relative PAH strengths for the BCD and SINGS 6.2 and 7.7μm fea-tures are virtually indistinguishable: 0.10 vs. 0.11 for 6.2μm, and 0.49 vs. 0.42for 7.7μm (Smith et al. 2007). However, for the longer wavelength PAHs at 8.6

L.K. Hunt et al.: PAHs and the ISM in Metal-Poor Starbursts 145

Fig. 1. PAH strengths relative to total PAH luminosity vs. total PAH power normalized

to total IR luminosity (TIR). The four panels show the the following dust features (DF):

top 8.6 μm, 11.3 μm, and 6.2 μm, 7.7 μm features. BCDs are plotted as filled (blue)

circles; open (green) squares correspond to SINGS HII nuclei; and open (red) triangles

to SINGS AGN (Smith et al. 2007). The horizontal dashed lines give the SINGS sample

medians (Smith et al. 2007), and the dotted ones the means for the BCD sample, taking

into account all objects with 7.7 μm detections. The vertical dotted line corresponds to

the BCD mean PAH power normalized to TIR.

and 11.3μm, the BCD fractions are >∼ 40% larger: 0.10 vs. 0.07 (8.6) and 0.17vs. 0.12 (11.3). Although with considerable scatter, the BCD mean PAH relativefractions for these features tend to exceed even the 10 to 90th percentile spreadsof the SINGS galaxies.

3.2 Profile widths and central wavelengths

Figure 2 shows the mean Drude profile widths and central wavelengths of sixaromatic features detected at >∼ 3σ in our sample; the numbers of BCDs with eachfeature are given in parentheses. The analogous quantities for the SINGS galaxiesare also plotted. In no case, were the PAHFIT profiles fixed to the SINGS galaxyparameters able to fit the BCD data; every BCD was better fitted by allowing theDrude profile widths (FWHMs) and central wavelengths to vary. In some cases,the differences between the χ2

ν with the SINGS “standard” profile parameters andthe best-fit parameters were small, <∼ 15%; but in others, the improvement of χ2

ν

was almost a factor of 2. The only systematic difference between the two samplesis the narrower profile width for the 8.6μm feature; only the very broadest BCDprofiles are as broad as those in more metal-rich systems.


Fig. 2. Median Drude pro-

file widths (FWHM) and cen-

tral wavelengths for seven aro-

matic features detected (�3σ) in

our sample. The error bars

are the standard deviations of

the measurements, and the num-

bers in parentheses are the BCDs

with significant detections. Open

squares correspond to SINGS

sample means (Smith et al. 2007).

3.3 Lack of small PAHs at low metallicity?

The larger relative intensities of the 8.6 and 11.3μm feature in BCDs and the nar-row profile width of the 8.6μm PAH could be due to a different size distributionof PAH populations at low metal abundances. Bauschlicher et al. (2008) find thatthe 8.6μm band arises from large PAHs, with Nmin

>∼ 100 carbon atoms. Hence,they suggest that the relative intensity of the 8.6μm band can be taken as an indi-cator of the relative amounts of large and small PAHs in a given population. Thelarge 8.6μm intensity relative to total PAH power in BCDs could be a signatureof fewer small PAHs (or more larger ones) at low metallicity. The relative lack ofsmall PAHs could also explain the low detection rate of the 6.2μm feature, sincemost of its intensity comes from PAHs with less than 100 C atoms (Schutte et al.1993; Hudgins et al. 2005). The size of a PAH molecule grows with increasingnumbers of carbon atoms (Draine & Li 2007). For a given maximum number ofcarbon atoms Nmax, a larger minimum number Nmin theoretically gives narrowerprofiles (Joblin et al. 1995; Verstraete et al. 2001). Although fitting intrinsicallyasymmetric PAH profiles by symmetric Drude profiles in PAHFIT is a simplifica-tion, the narrower width of the 8.6μm feature could be a signature of relativelylarger PAHs at low metallicity. The difference disappears at 7.6μm and is lesspronounced at longer wavelengths than 8.6μm because both small and large PAHsizes contribute to these bands1 (Schutte et al. 1993; Bauschlicher et al. 2009).

Another indication that low-metallicity BCDs may be lacking the smallest sizePAHs comes from a correlation analysis. In a detailed study of Galactic Hii regions,young stellar objects, reflection and planetary nebulae (RNe, PNe), and evolvedstars, Hony et al. (2001) found only a few correlations among PAH features. Oneof these is between 6.2 and the 12.7μm emission, which is shown for our sample ofBCDs in Figure 3; only galaxies with >∼ 3σ detections in at least two features areplotted. Most galaxies with 6.2μm PAH detections and 12+log(O/H)≥8.1 (filledcircles) are similar to the Hii regions studied by Hony et al. (2001). On the other

1The 7.8 μm band may be dominated by larger PAHs only (Bauschlicher et al. 2009); butthis band is present in only two of the BCDs in our sample, Haro 3 and II Zw 70.

L.K. Hunt et al.: PAHs and the ISM in Metal-Poor Starbursts 147

Fig. 3. The integrated flux in

the PAH features at 6.2 μm,

11.3 μm, and 12.7 μm, plotted

in two sets of ratios: 6.2/11.3

vs. 12.7/11.3. Only those BCDs

with at least one ratio with

11.3 μm are shown: filled (blue)

circles correspond to those BCDs

with 12+log(O/H)≥8.1; open

ones to those objects with lower

metallicities. The other data are

taken from Hony et al. (2001)

and correspond to PNe, RNe,

intermediate-mass star-forming

regions and Hii regions. The

SINGS average is shown as a

×, with error bars reporting

standard deviations over the

sample (Smith et al. 2007).

hand, BCDs with lower O/H occupy a region of the plot with small 6.2/11.3 ratios.Although the statistics are poor, small 6.2/11.3 ratios are found for most of thelowest-metallicity BCDs. Because the 11.3μm feature arises from non-adjacent or“solo” CH groups, dominant 11.3μm emission, like 8.6μm emission, implies largePAH species, with �100–200 carbon atoms and long straight edges (Hony et al.2001; Bauschlicher et al. 2008; Bauschlicher et al. 2009). Moreover, fewer smallPAHs reduce the intensity of the 6.2μm emission, as the bulk of this feature comesfrom PAHs containing less than 100 C atoms (Schutte et al. 1993; Hudgins et al.2005). Therefore, small 6.2/11.3μm band ratios at low metallicities could implyan overall deficit of small PAHs.

A third indication that there is a deficit of small PAHs in a metal-poor ISMis the intensity of the 8.6μm feature relative to the 7.7μm blend. As men-tioned above, Bauschlicher et al. (2008) find that 8.6μm band arises from largePAHs, with Nmin

>∼ 100 carbon atoms. On the other hand, the 7.7μm featurecontains emission from both small and large PAH sizes (Schutte et al. 1993;van Diedenhoven et al. 2004). For the 9 BCDs with the 7.7 and 8.6μm PAHfeatures detected at >∼ 3σ, the mean 8.6/7.7 flux ratio is 0.48, and the medianis 0.33. This ratio is at least double that in the more metal-rich SINGS galaxies,with 8.6/7.7 ∼0.18 in the mean, and a range of 0.11−0.21 (Smith et al. 2007); 7 of9 BCDs are outside the upper 90th percentile limit of 0.21. Again, although thenumber statistics are small, the large 8.6/7.7μm flux ratios of the PAHs in theselow-metallicity BCDs appear to be dominated by the largest PAH species.


4 Conclusions

In summary, there are three different lines of evidence which tentatively suggestthat the PAH populations in low-metallicity BCDs are deficient in the smallestsizes (Nmin

<∼ 50 C atoms). The IRS spectra show flux ratios typical of the largestPAHs modeled so far, with Nmin

>∼ 100 C atoms. Correlations within our samplesuggest that the main trend is with radiation field intensity and hardness, ratherthan with metallicity. Apparently, the ISM environment in the BCD Hii regionsis too harsh for the smallest PAHs, possibly destroying them through sputtering.Perhaps only the PAH population with the largest species can survive the extremephysical conditions in the low-metallicity Hii regions in the BCDs.

We would like to thank the organizers for a very stimulating and well-organized conference.

References

Bauschlicher, C.W., Jr., Peeters, E., & Allamandola, L.J., 2008, ApJ, 678, 316


Brandl, B. R., et al., 2006, ApJ, 653, 1129

Draine, B. T., & Li, A., 2007, ApJ, 657, 810


Galliano, F., Madden, S.C., Jones, A.P., Wilson, C.D., & Bernard, J.-P., 2005, A&A,434, 867


Houck, J.R., et al., 2004a, ApJS, 154, 18

Hudgins, D.M., Bauschlicher, C.W., Jr., & Allamandola, L.J., 2005, ApJ, 632, 316

Hunt, L.K., Thuan, T.X., Izotov, Y.I., & Sauvage, M., 2010, ApJ, 712, 164

Joblin, C., Boissel, P., Leger, A., D’Hendecourt, L., & Defourneau, D., 1995, A&A, 299,835

Kennicutt, R.C., Jr., et al., 2003, PASP, 115, 928



Peeters, E., Spoon, H.W.W., & Tielens, A.G.G.M., 2004a, ApJ, 613, 986



Smith, J.D.T., et al., 2007, ApJ, 656, 770

Taylor, C.L., Kobulnicky, H.A., & Skillman, E.D., 1998, AJ, 116, 2746


Verstraete, L., et al., 2001, A&A, 372, 981

Wu, Y., Charmandaris, V., Hao, L., et al. 2006, ApJ, 639, 157


INTRODUCTION TO AMUSES: AKARI SURVEY WITH AWINDOW OF OPPORTUNITY

J.H. Kim1, M. Im1, H.M. Lee2, M.G. Lee2 and the AMUSES team

Abstract. With advancement of infrared space telescopes during thepast decade, infrared wavelength regime has been a focal point tostudy various properties of galaxies with respect to evolution of galax-ies. Polycyclic Aromatic Hydrocarbons (PAHs) have emerged as oneof the most important features since these features dominate the mid-infrared spectra of galaxies. These PAH features provide a great handleto calibrate star formation rates and diagnose ionized states of grains.However, the PAH 3.3 μm feature has not been studied as much asother PAH features since it is weaker than others and resides outsideof Spitzer capability, although it will be the only PAH feature accessibleby JWST for high-z galaxies. AKARI mJy Unbiased Survey of Extra-galactic Sources in 5MUSES (AMUSES) intends to take advantage ofAKARI capability of spectroscopy in the 2 ∼ 5 μm to provide an un-biased library of 44 sample galaxies selected from a parent sample of5MUSES, one of Spitzer legacy projects. For these 3.6 μm flux limitedsample galaxies whose redshifts range between 0 < z < 1, AMUSESwill calibrate PAH 3.3 μm as a star formation rate (SFR) indicatorwhile measuring ratios between PAH features. We present preliminaryresults of AMUSES.

1 Introduction

Infrared (IR) astronomy has progressed tremendously over the past decades thanksto various space missions, such as the Infra-Red Astronomical Satellite (IRAS)(Soifer et al. 1987), the Infrared Space Observatory (ISO) (Kessler et al. 1996;Genzel & Cesarsky 2000), and the Spitzer space telescope (Werner et al. 2004).While providing clues to numerous astrophysical subjects, these space-based ob-servatories also improve extragalactic astronomy since these IR wavelengths carry

1 CEOU, Department of Physics and Astronomy, Seoul National University2 Department of Physics and Astronomy, Seoul National University



a vast information on galaxies. For example, the shortest IR wavelengths (upto 10 μm) represent the photospheric light and are good stellar mass indicators(Regan & The Sings Team 2004). Then, dust grains at different temperaturesare represented by different ranges of IR wavelengths (Genzel & Cesarsky 2000).IR wavelengths also carry numerous lines from hydrogen molecule and gas-phasespecies. However, the most important aspect of IR astronomy on galaxies is prob-ably that IR emission represents dust-obscured star formation activity of galaxies(Genzel & Cesarsky 2000). While other shorter wavelength star formation rate(SFR) proxies, such as hydrogen recombination lines and ultraviolet (UV) emis-sions rely on the measurement of photoionizing UV photons from heavy stars and,thus, suffer from extinction, the bolometric IR luminosity measures dust-obscuredstar formation within galaxies and is less affected by extinction. However using thebolometric IR luminosity as a SF indicator has two caveats. First, not only newlyformed heavy stars, but also evolved stellar populations can heat dust componentswithin galaxies (cf. Calzetti elsewhere in this volume). Then it is extremely trickyto understand the whole range of IR spectral energy distribution (SED) for high-zgalaxies.

Therefore, many studies have attempted calibrate reliable IR SF proxies tobolometric IR luminosities in recent years. Among these IR SFR proxies, poly-cyclic aromatic hydrocarbons (PAHs) have gotten enormous attention due to theirubiquity and strong potential as diagnostics of other properties. On the basis ofthe aromatic IR band emission, PAHs are considered to be present in a wide rangeof objects and environments, such as post-AGB stars, planetary nebulae, HII re-gions, reflection nebulae and the diffuse interstellar medium (Puget et al. 1985;Allamandola et al. 1989). The PAH mid-IR features are believed to contributeup to 10% of the total IR luminosity of star forming galaxies (Helou et al. 2001;Peeters et al. 2002; Smith et al. 2007).

Numerous recent studies measure PAH band fluxes and equivalent widths(EWs) in oder to calibrate these emission features as SFR proxies within theGalactic environments and galaxies at higher redshift. These studies reveal thatthere exist differences in PAH EWs and LPAH/LIR ratios between local valuesand high redshift ones (Helou et al. 2001; Peeters et al. 2002; Smith et al. 2007).Since PAH band ratios reflect variations in physical conditions within environ-ments, such as ionization states of dust grains and metallicity (Smith et al. 2007;Galliano et al. 2008; Gordon et al. 2008. See papers by Draine and by Gallianoelsewhere in this volume), more detailed study on this subject will put a betterconstraint on physical conditions of PAH emission sites and calibration of PAHbands as SFR proxies.

Then, most studies on PAH emission features concentrate on stronger bands,such as 6.2, 7.7, and 11.3 μm due to the relatively weaker strength of 3.3 μmfeature and lack of wavelength coverage by instruments on board Spitzer. TheJapanese space-based IR observatory, AKARI can fill this gap (Murakami et al.2007). Launched on February 26th, 2006, AKARI has a 68.5 cm telescope whichis cooled down to 65 K. It has two instruments, Far-Infrared Surveyor (FIS) andInfraRed Camera (IRC), whose wavelength coverages range from 1.8 to 180 μm.

J.H. Kim et al.: Introduction to AMUSES 151

Developed to carry out all-sky survey, AKARI has two operation modes, surveymode and pointing mode. More details on AKARI, its operation and accomplish-ments can be found in the literature. Spectroscopic observations of PAHs havealso been reported by Onaka et al. elsewhere in this volume.

We present one of AKARI mission projects (MPs), AKARI mJy UnbiasedSurvey of Extragalactic Sources (AMUSES). The main scientific goal of AMUSESis to construct a continuous spectral library over the wavelength window between2.5 and 40 μm for a subsample of 5 mJy Unbiased Spitzer Extragalactic Survey(5MUSES) (Wu et al., accepted). AMUSES enables to access the 2.5 – 5 μm spec-tral window which includes the 3.3 μm emission feature. Moreover, the 3.3 μmPAH feature is the only dust emission feature at high redshift (z > 4.5) accessi-ble to JWST. Therefore, understanding and calibrating the 3.3 μm feature withrespect to the total IR luminosity, the SFR and other PAH emission features forlocal galaxies is critical for interpreting JWST spectra of dusty galaxies in future.

2 Sample selection

We select our sample from 5MUSES. 5MUSES, one of the Spitzer Legacy surveys,performs a mid-infrared spectroscopic observation of extragalactic sources brighterthan 5 mJy at 24 μm in the Spitzer First Look Survey (FLS) field and four subfieldsof Spitzer Wide-area Infrared Extragalactic (SWIRE) survey with the InfraredSpectrograph (IRS) onboard the Spitzer space telescope. The main scientific goalof 5MUSES is to provide an unbiased library of infrared spectra from 5 to 40 μmof sources which have not been sought after in previous studies. Since the mainobjective of AMUSES is to detect the 3.3 μm PAH feature, we narrow down oursample to be brighter than 1 mJy at 3.6 μm. Based on the Spitzer IRAC 3.6 μmdata for the 5MUSES sample, we find that 60 of the 330 5MUSES galaxies satisfythe flux cut of 1mJy at 3.6 μm. In order to detect the 3.3 μm feature, we limit theredshift range to z < 0.5. With this additional cut, we narrow down the samplesize to 51 galaxies. In addition to this base sample, we add 10 targets with their3.6 μm flux brighter than 0.7 mJy whose redshifts could not be determined byoptical spectroscopy. After this selection, 60 target galaxies remain.

Then 44 targets are approved for the final program. Based on the SED clas-sification, there are 22 AGNs and 17 starburst galaxies. The rest of the sampleshows a composite SEDs. The redshift distribution is presented in Figure 1.

3 Observation and data reduction

The main goal of AMUSES is to detect the PAH 3.3 μm feature with S/N > 5with IRC (Onaka et al. 2007). IRC is a workhorse for AKARI and has a fieldof view of roughly 10′ × 10′. Pixel scales for three independent cameras rangefrom 1.45′′ through 2.5′′. Each observation during the pointing mode has a max-imum exposure time of ten minutes, but actual exposure times run between sixto seven minutes. In order to achieve this S/N for sources with 1 mJy at 3.6 μm,


Fig. 1. Redshift distribution of the sample of galaxies of AMUSES plotted against 3.6 μm

flux. Symbols represent the SED classification of target galaxies based on their MIR colors

by 5MUSES collaboration. Open squares represent the AGN-type SED, while crossess

represent the starburst-type SED. Open circles represent the composite SED. AGN-type

targets are more evenly distributed across the redshift range, while starburst-type targets

are more clustered at lower redshifts.

we decided to have three pointings for each target with NIR grism (NG) whileadding one pointing with NIR prism (NP) in order to improve continuum extrac-tion. These observations are carried out with a slit aperture whose dimensionis 1′ by 1′. In total, 51 pointings are observed for 20 target galaxies. Amongthem, ten sources have completed their scheduled observations. Data reduction isperformed with the IRC spectroscopy pipeline1. Additional cosmic ray removal,stacking multipointing exposures, and sigma clipping during stacking are executedindividually after running the IRC spectroscopy pipeline. Final one-dimensionalspectra are extracted from two-dimensional spectral images which are binned bythree pixels along the wavelength direction.

4 Results

We present several sample spectra in Figure 2. Overall, the 3.3 μm PAH emissionfeature is pretty weak across the sample even if it is detected. However, it isstill noticeable that its strength is stronger for galaxies with starburst SED classthan for AGN class. Most galaxies in which the 3.3 μm PAH feature is detectedwith a significant S/N are classified as starburst galaxies. The only exceptions areVIIZw353 and 2MASX J16205879+5425127 in which the 3.3 μm PAH detectionis marginal. The latter is also classified as composite SED.

1http://www.ir.isas.jaxa.jp/ASTRO-F/Observation/DataReduction/IRC/

J.H. Kim et al.: Introduction to AMUSES 153

Fig. 2. Sample spectra of AMUSES based on IRC and IRS spectra. Notice the gaps

between two instruments around 5 μm. All spectra are deredshifted. Vertical dashed

lines indicate the most notable PAH emission features at 3.3, 6.2, 7.7, 11.2, and 12.7 μm.

The two galaxies on the left panels are classified as starburst galaxies based on their SEDs

while the two galaxies on the right panels are classified as AGN, or have a composite

SED. The 3.3 μm PAH feature is always the weakest and difficult to detect compared to

the other prominent PAH emission features.

We also present a couple of plots showing correlations between EWs of the3.3 μm, and 6.2 μm features (Fig. 3). Bigger symbols represent samples ofAMUSES. Asterisks represent galaxies with starburst SEDs while filled circlesrepresent galaxies with AGN SEDs. A diamond represent a galaxy with compos-ite SED. Overplotted are the data from Imanishi et al. (2008). Asterisks representtheir samples classified as HII-like galaxies, i.e. star-forming galaxies, while filledcircles represent AGNs. Within the sample of Imanishi et al. (2008), two SEDclasses do not differ much in correlations between EWs of the 3.3 μm, and 6.2 μmbands. They are relatively well matched, albeit with big scatter. AMUSES sam-ple galaxies, most of which are starburst galaxies, have even bigger scatter whilegenerally following the trend set by the sample of Imanishi et al.

5 Summary

We introduce a mission project of AKARI, AMUSES and its preliminary results.With AMUSES, we investigate the 3.3 μm PAH feature and its correlation withother PAH features as well as other properties of sample galaxies while constructingan unbiased library of sample galaxies selected from a parent sample of 5MUSES.So far, we have measured fluxes and EW of the 3.3 μm PAH feature for 20 targetgalaxies and have compared their EW to the Imanishi et al. (2008) sample.


Fig. 3. Correlation between equivalent widths of the 3.3 μm PAH line and the 6.2 μm

PAH line. Asterisks represent galaxies with starburst SEDs while filled circles represent

galaxies with AGN SEDs. A diamond represent a galaxy with composite SED. Over-

plotted are the data from Imanishi et al. (2008). Their AGN samples are represented by

small filled cicles and their HII-like samples are represented by small asterisks.

This work was supported by the National Research Foundation of Korea(NRF) grant funded bythe Korean government(MEST), No. 2009-0063616. This research is based on observations withAKARI, a JAXA project with the participation of ESA.

References


Calzetti, D., et al., 2007, ApJ, 666, 870


Genzel, R., & Cesarsky, C.J., 2000, ARA&A, 38, 761


Helou, G., Malhotra, S., Hollenbach, D.J., Dale, D.A., & Contursi, A., 2001, ApJ, 548,L73

Imanishi, M., Nakagawa, T., Ohyama, Y., et al., 2008, PASJ, 60, 489

Kessler, M.F., et al., 1996, A&A, 315, L27

Murakami, H., et al., 2007, PASJ, 59, 369

Onaka, T., et al., 2007, PASJ, 59, 401


Puget, J.L., Leger, A., & Boulanger, F., 1985, A&A, 142, L19

Regan, M., & The Sings Team 2004, 35th COSPAR Scientific Assembly, 35, 4098

Smith, J.D.T., et al., 2007, ApJ, 656, 770

Soifer, B.T., Neugebauer, G., & Houck, J.R., 1987, ARA&A, 25, 187

Werner, M.W., et al., 2004, ApJS, 154, 1

Wu, Y., et al., 2010, ApJ, accepted

The Lifecycle

of PAHs in Space


PAH EVOLUTION IN THE HARSH ENVIRONMENTOF THE ISM

H. Kaneda1, T. Onaka2, I. Sakon2, D. Ishihara1, A. Mouri1,M. Yamagishi1 and A. Yasuda1

Abstract. We review observational results of PAH emission in harshinterstellar environments, which are mostly based on recent works bySpitzer and AKARI. The harsh environments include shock regions inour Galaxy, the ionized superwinds and haloes of external galaxies,and the hot plasmas of elliptical galaxies. Owing to the unprecedentedhigh sensitivity for PAH emission with Spitzer and AKARI, it is foundthat an appreciable amount of PAHs are present in some cases withsuch hostile conditions. Some of them exhibit unusual PAH interbandstrength ratios, reflecting either evolution of PAHs or physical condi-tions of the ISM. The distribution of the PAH emission, as comparedto that of dust emission, is shown to discuss their ways of evolutionand survival.

1 Introduction

The evolution and survival of polycyclic aromatic hydrocarbons (PAHs) in theharsh environments of the interstellar medium (ISM) are interesting and impor-tant problems to be addressed for understanding the lifecycle of PAHs in space.PAHs are expected to be processed and destroyed quite easily in interstellar shockregions and hot plasmas (e.g. Tielens 2008). In general, as observed in our Galaxyand nearby galaxies, most PAHs are associated with neutral gas and their emissionis weak in ionized regions with strong radiation field probably due to destruction(e.g. Boulanger et al. 1988; Desert et al. 1990; Bendo et al. 2008). However thetrue destruction efficiency of PAHs has not yet been well understood from astro-nomical observations because of scarcity of relevant data and their faintness in themid-IR emission in such conditions, although there are several theoretical studies(Jones et al. 1996; Micellota et al. 2010a, 2010b). Now the situation has been

1 G. School of Science, Nagoya University, Chikusa-ku, Nagoya 464-8602, Japan2 Graduate School of Science, The University of Tokyo, Bunkyo-ku, Tokyo 113-0033, Japan



improved with the advent of Spitzer and AKARI; owing to their unprecedentedhigh sensitivity for PAH emission, it is found that PAHs are indeed destroyed quiteefficiently as compared to dust grains in most objects with such hostile conditions,but in some cases, an appreciable amount of PAHs are still present. Some of themexhibit unusual PAH interband strength ratios, reflecting either evolution of PAHsor physical conditions of the ISM.

From recent theoretical studies combined with laboratory data, Micellota et al.(2010a, 2010b) calculated the collisional destruction efficiency of PAHs in inter-stellar shocks and hot plasma. According to their results, for shock velocities lowerthan 100 km s−1, PAHs in shocks are not completely destroyed but their struc-tures are significantly affected. In such a slow shock, PAHs may even be formed byshattering of carbonaceous grains. On the other hand, for shock velocities higherthan 200 km s−1, PAHs are completely destroyed in postshock hot plasma.

In hot plasmas with temperatures lower than 3 × 104 K, ion nuclear collisionsdominate in the destruction process, which destroy PAHs not so efficiently; forexample, in a hot plasma with the density and temperature of 104 cm−3 and104 K (e.g. Orion), small (50–200 C-atom) PAHs can survive against the colli-sional destruction for a time as long as 107 years. For hotter plasma, however,electron collisions dominate in the destruction process, which destroy PAHs veryefficiently. For example, in a hot plasma with the density and temperature of10−2 cm−3 and 107 K (e.g. M 82 superwinds), small PAHs can survive only for100 years agaist collisional destruction (Micellota et al. 2010b). It should benoted that these lifetimes of PAHs are two to three orders of magnitude shorterthan those for dust grains with the same size, because the sputtering yields of3-dimensional dust grains are much smaller than unity, while the dissociationyields of 2-dimentional PAHs are close to unity. In HII regions with strong UVradiation field, the photo-dissociation of PAHs is more important, where the life-times of PAHs are ∼105 year for typical HII regions while they are ∼108 years forthe diffuse ISM (Allain et al. 1996).

Below, we review observational results of PAH emissions in harsh interstellarenvironments, which are mostly based on recent works by Spitzer and AKARI. Theharsh environments include shock around SNRs and intense UV field in a massivecluster in our Galaxy, the ionized superwinds and haloes of external galaxies, andthe hot plasmas of elliptical galaxies.

2 PAHs in harsh environments in our Galaxy

According to Spitzer studies of PAH emission from Galactic SNRs by Reach et al.(2006), only 4 SNRs, out of 95, indicate the presence of PAH emission from IRACcolors. One example is 3C396, where the 8 μm PAH map shows the presence offilamentary structures inside the radio shell. The Spitzer IRS also detected PAHemission features from the spectra of the SNRs W28, W44, and 3C 391 (Neufeldet al. 2007), but their association with the SNRs was not so clear. For theSNR N132D in the Large Magellanic Cloud (LMC), a PAH 17 μm broad featurewas detected from Spitzer/IRS spectra of the southern rim across the shock front

H. Kaneda et al.: PAHs in Harsh Environments 159

Fig. 1. AKARI 9 μm and 18 μm images of the Tycho SNR on the left and right of the

upper panel, respectively (Ishihara et al. 2010), and the AKARI 9 μm large-scale image

around the Vela SNR in the lower panel, which are created from the AKARI mid-IR

all-sky survey data. The ROSAT X-ray contours are superposed on the Vela SNR image.

(Tappe et al. 2006). The 17 μm broad feature is interpreted as the emission oflarger PAHs with a few thousand carbon atoms in C–C–C in- and out-of-planebending modes (Van Kerckhoven et al. 2000; Peeters et al. 2004). The lifetime oflarge (4000 C-atom) PAHs is estimated to be ∼200 years (Micellota et al. 2010b)while the age of the SNR is ∼2500 years, and therefore such large PAHs may havebeen recently swept up by the blast wave and not yet completely destroyed by theshock (Tappe et al. 2006).

AKARI mid-infrared (IR) all-sky survey data are useful to discuss the presenceof PAHs in SNRs. They consist in photometry in two wide bands centered atwavelengths of 9 μm and 18 μm with effective wavelength ranges of 6.7–11.6 μmand 13.9–25.6 μm, respectively (Onaka et al. 2007). The 9 μm band includes thePAH features at 7.7, 8.6, and 11.3 μm, while the 18 μm band traces warm dustcontinuum emission. The AKARI 9 μm band is relatively broad as compared tothe MSX and Spitzer bands for PAH emission, and it is therefore more likely tobe dominated by PAH features and less affected by emission lines.

Structures associated with SNRs are often identified in the AKARI 18 μmband, whereas they are not in the AKARI 9 μm. For example, the upper panel ofFigure 1 shows the AKARI 9 μm and 18 μm images of the Tycho SNR at an age


of ∼440 years, where a spherical shell structure is clearly seen in the warm dustcontinuum emission while it is hardly seen in the PAH emission. The dust emissionshows a beautiful spatial correlation with the X-ray and CO emission (Ishiharaet al. 2010). With the current shock speeds (∼3000 km s−1; Reynoso et al. 1997)and the thickness of the dust emission region, we estimate dust residence time inpostshock plasma to be ∼50 years, where the X-ray plasma has temperature andelectron density of 8 × 106 K and 10 cm−3, respectively (Warren et al. 2005).Then Figure 1 demonstrates that the dust can survive while the PAHs are almosttotally destroyed in the 50 year timescale in a hot plasma with the above plasmaparameters, which is consistent with the theoretical prediction. The lower panelof Figure 1 shows an AKARI 9 μm image around the Vela SNR (∼1.1× 104 yearsold), on which the X-ray contours are superposed. In this large-scale image, wecan clearly see an anti-correlation between the PAH and the X-ray emission, againdemonstrating effective destruction of PAHs in the postshock hot plasma.

Other than SNRs, AKARI also reveals the properties of PAHs in harsh envi-ronments of clouds with energetic phenomena with the help of large-scale CO databy the NANTEN telescope. One example is RCW49, Wd2, one of the most mas-sive star clusters in our Galaxy. The massive cluster formation was likely triggeredby cloud-cloud collision ∼4 × 106 years ago because the NANTEN CO observa-tions revealed two large molecular clouds with a systematic difference in the speed(∼10 km s−1), the crossing point of which exactly corresponds to the position ofWd2 (Furukawa et al. 2009). Figure 2 shows the AKARI 9 μm and 18 μm com-posite map of RCW49, where the central region is saturated in the 18 μm band.In the figure, two examples of the AKARI near-IR 2–5 μm slit spectra are showntogether with their observational positions, from each of which a 3.3 μm emissionis detected. However we should be careful about this interpretation because the3.3 μm emission is also caused by the hydrogen recombination Pfδ line at 3.30 μmoverlapping with the PAH 3.3 μm band. Because the intensity of the Pfδ lineshould be ∼0.7×Pfγ for the case B condition, the 3.3 μm emission is most likelydominated by the PAH emission only if the 3.3 μm line is much stronger than thePfβ and Pfγ lines at 4.66 μm and 3.75 μm, respectively. By taking this into ac-count, the presence of the PAH 3.3 μm emission with its subfeatures at 3.4–3.5 μmis confirmed in many regions for which spectroscopic observations have been per-formed around Wd2 except for its very central region. This physically reflects thepresence of very small PAHs because PAHs with the number of C-atoms >100 orthe size >6 A cannot emit the 3.3 μm feature efficiently (Draine & Li 2007). Itis thus found that even very small PAHs survive for ∼4× 106 years under intenseUV radiation field near very active star-forming region, which shows a strikingcontrast to the SNRs where PAHs are very easily destroyed.

In general, suppressed PAH emission relative to dust emission (e.g. unusuallylow ratios of the AKARI 9 μm to 18 μm band intensity) may serve as evidence ofshock gas heating rather than radiative heating. For example, the NANTEN COobservations by Fukui et al. (2006) revealed the presence of large-scale molecularloops in the Galactic center region. The theoretical simulation by Machida et al.(2009) predicted magnetic loops can buoyantly rise due to the Parker instability


Fig. 2. AKARI 9 μm (blue) and 18 μm (red) composite image of the RCW49 region,

where the center is saturated in the 18 μm band. Two examples of AKARI near-IR 2–

5 μm slit spectra are shown together with their observational positions.

with disk gas, which cools at the top and then slides down along the loop witha speed of ∼30 km s−1, presumably resulting in shock gas heating. If this is thecase, we can expect that PAH emission is relatively suppressed at the foot pointsof the molecular loops.

3 PAHs in the haloes of external galaxies

Several authors have reported the detection of PAH emission in the haloes ofnearby edge-on galaxies. As for relatively quiescent normal galaxies, PAH emis-sions in the haloes of NGC 5529 and NGC 5907 extending up to ∼10 kpc and∼6 kpc above the disks were detected with ISO by Irwin et al. (2007) and Irwin& Madden (2006), respectively. The latter was the first report of the detection ofPAHs in the halo of an external galaxy. For NGC 5907, Irwin & Madden (2006)found that the ratio of the PAH 11.3 to 7.7 band increases with distance fromthe galactic plane and also pointed out that there is a similarity in the spatialdistribution between the PAH 6.3 μm and Hα emission. Whaley et al. (2009)showed the presence of PAH emission in the halo of NGC 891 extending up to∼3 kpc above the disk with Spitzer and ISO. They compared the vertical extentsof various components and found that the cool dust extends to larger scale heightsthan the PAHs and warm dust.

For more active galaxies, PAH emission from the galactic superwind of M 82was observed with Spitzer and was found to widely extend around the halo up to


Fig. 3. AKARI 7 μm band contour map of M82 overlaid on the continuum-subtracted

Hα image (Kaneda et al. 2010a) on the left, and H2 2.12 μm emission image (Veilleux

et al. 2009) on the right of the upper panel. In the lower panel, the low-level 7 μm,

15 μm, and 160 μm band images of M82 including the Cap region, overlaid on the

XMM/Newton X-ray (0.2–10 keV) contour map (Kaneda et al. 2010a).

∼6 kpc from the center (Engelbracht et al. 2006; Beirao et al. 2008). Howeverone serious problem is that the central starburst core is dazzlingly bright due totremendous star-forming activity in the central region of M 82 (Telesco & Harper1980); indeed M 82 is the brightest galaxy in the mid- and far-IR after the Magel-lanic Clouds. Then, instrumental effects caused by saturation made quantitativediscussion about the low-surface brightness distribution around the halo ratherdifficult. Tacconi-Garman et al. (2005) performed high-spatial observations ofthe central region of NGC 253 with the IR camera and spectrograph ISAAC onESO Very Large Telescope UT1. They found the presence of PAH emission inthe galactic superwind of NGC 253 extended on a 0.1 kpc scale on the basis of the3.28 μm narrow (0.05 μm width) band image, for which, however, the contributionby the Pfδ line may have to be taken into account as mentioned above.

AKARI also observed M 82 over a wide area from the center to the halo inthe mid- and far-IR in deep pointed observations (Kaneda et al. 2010a). Oneof the AKARI merits is a large dynamic range of signal detection; AKARI has


a special mode to observe bright sources so that the instruments may not suffersaturation effects. The AKARI M 82 data consist of four mid-IR narrow-bandimages at reference wavelengths of 7 μm (effective band width: 1.75 μm), 11 μm(4.12 μm), 15 μm (5.98 μm), and 24 μm (5.34 μm), the allocation of which is idealto discriminate between the PAH emission features (7 μm band tracing the PAH7.7 and 8.6 μm features, 11 μm band tracing the 11.3 and 12.7 μm features) andthe MIR dust continuum emission (15 μm, 24 μm). The data also include fourfar-IR photometric bands; 2 wide bands at central wavelengths of 90 μm (effectiveband width: 37.9 μm) and 140 μm (52.4 μm) and 2 narrow bands at 65 μm(21.7 μm) and 160 μm (34.1 μm). The left upper panel of Figure 3 shows theAKARI 7 μm band contour map of M 82 for a 10×10 arcmin2 area, which revealsfilamentary structures widely and faintly extended into the halo. As compared tothe Hα distribution in color, a very tight correlation is found between the PAHsand the ionized superwind, providing evidence that the PAHs are well mixed in theionized superwind gas and outflowing from the disk. In contrast, the H2 2.12 μmemission (Veilleux et al. 2009) shows a relatively loose correlation with the PAHemission especially in the northern halo (the right upper panel of Fig. 3). Thedeprojected outflow velocity of the Hα filaments is 525− 655 km s−1 (Shopbell &Bland-Hawthorn 1998). Therefore the PAHs seem to have survived in such a harshenvironment for about 5×106 years to reach the observed positions, ∼3 kpc abovethe disk, in the ionized superwind but not X-ray-emitting hot plasma because thedestruction time in a hot plasma should be very short, ∼100 years (Micelotta et al.2010b).

Far beyond the disk of M 82, there is a region called X-ray Cap. The lowerpanel of Figure 3 shows the AKARI wide-area images of M 82 covering the Capregion at 7 μm, 15 μm, and 160 μm, exhibiting the low-brightness distributionsof the PAH, warm dust, and cool dust emission, respectively. The X-ray contoursare overlaid for each image. Hence PAH emission is widely extended in the Capregion as well, which is presumably caused by the past tidal interaction of M 82with M 81 about 108 years ago (Walter et al. 2002). There is some anticorrelationbetween the X-ray and PAH emission at the Cap position and also at the upstreamof the X-ray superwind. The 15 μm warm dust emission is reasonably weak farbeyond the disk, and the 160 μm band clearly shows emission extended to thesame direction as the X-ray superwind. Therefore the cool dust is also outflowing,entrained by the hot plasma. By comparing these maps, it is found that thedestruction of PAHs in the X-ray plasma is clearly faster than the destruction offar-IR dust.

Another clear example of PAH detection from galactic superwinds is obtainedfor NGC 1569, a starburst dwarf galaxy possessing a very young (starburst peaked∼5×106 years ago) and low-metalicity (Z ∼ 1/4 Z�) environment; thus we expecta very small contribution from old stars to interstellar PAHs (Galliano et al. 2008).The background-subtracted spectra obtained by AKARI revealed the presence ofthe 3.3, 6.2, 7.7, and 11.3 μm PAH emission features in the prominent X-ray/Hαfilament produced by the galactic superwinds of NGC 1569 (Onaka et al. 2010).The filament age is approximately 106 years, whereas the destruction timescale


NGC2974

5 40

8

1 10

Sur

face

Brig

htne

ss (

MJy

/sr)

Wavelength (um)

NGC3962

5 40

1

10

Sur

face

Brig

htne

ss (

MJy

/sr)

Wavelength (um)

Fig. 4. Mid-IR spectra of the centers of the elliptical galaxies NGC2974 and NGC3962

obtained by the Spitzer/IRS staring observations. The figures are taken from Kaneda

et al. (2008).

of small PAHs must be shorter than 100 years. To reconcile the discrepancy, weexpect that PAHs may be produced by fragmentation of carbonaceous grains inshocks (Onaka et al. 2010).

4 PAHs in elliptical galaxies

Elliptical galaxies provide a unique environment: little UV from old stars andinterstellar space filled by hot plasma, which are hostile conditions for PAH emis-sion. Nevertheless, the early study with ISO by Xilouris et al. (2004) showed ahint of PAH emission in the spectral energy distributions of 10 elliptical galaxiesout of 18. Then Spitzer/IRS clearly detected PAH emissions from the spectra of14 elliptical galaxies out of 18 galaxies (Kaneda et al. 2008). Figure 4 shows suchexamples, where the prominent 11.3 μm and the broad 17 μm feature can be seenwhile usually the strongest 7.7 μm feature is rather weak. The relative weaknessof the 7.7 μm feature is almost commonly observed among the elliptical galaxies.Since the 11.3 μm feature is attributed to a C-H out-of-plane bending mode mainlyarising from neutral PAHs (e.g. Allamandola et al. 1989), a natural interpretationis that neutral PAHs are dominant over ionized PAHs due to soft radiation fieldfrom old stars, which is typical of elliptical galaxies. As another case supportingthis interpretation, from a systematic study of mid-IR spectra of Galactic regions,Magellanic HII regions, and galaxies of various types (dwarf, spiral, starburst) byISO and Spitzer, Galliano et al. (2008) found that the 6.2, 7.7, and 8.6 μm bandsare essentially tied together, while the ratios between these bands and the 11.3 μmband vary by one order of magnitude. They concluded that the relative variationsof the band ratios are mainly controlled by the fraction of ionized PAHs.

Alternatively, the characteristics of the PAH emission may be regulated bynuclear activities of the galaxies. A significant fraction of local elliptical galaxiesare known to harbor a low-luminosityAGN(LLAGN), including a low-ionizationnuclear emission region (LINER) nucleus (e.g. Ho et al. 1997). Smith et al. (2007)


Fig. 5. PAH 7.7 μm/11.3 μm ratios plotted against the line ratios of [NeIII]/[NeII] taken

from Smith et al. (2007) for the galaxies with HII-dominated nuclei and LLAGNs, and

from Kaneda et al. (2008) for the elliptical galaxies.

have reported that the peculiar PAH features with an unusually low ratio of the7.7 μm to the 11.3 μm emission strength arise from systems with relatively hardradiation fields of LLAGNs. Figure 5 shows a plot of PAH 7.7 μm/11.3 μm versus[NeIII]/[NeII] taken from Smith et al. (2007) and the elliptical galaxy data fromKaneda et al. (2008) are plotted together. Generally, the [NeIII]/[NeII] is anindicator of radiation hardness, and the elliptical galaxy data seem to follow aglobal trend. If this trend is true, PAHs in the elliptical galaxies may be affectedby hard radiation from the LINER nuclei through selective destruction of smallerPAHs. Indeed, similar weakness of PAH features in the 5–10 μm range relative tostrong 11.3 μm features were also obtained for radio galaxies (Leipski et al. 2009)and IR-faint LINERs (Sturm et al. 2006).

Recent deep spectral mapping with Spitzer/IRS clearly reveals the spatial dis-tributions of PAH emission in elliptical galaxies for the first time (Kaneda et al.2010b). An exemplary case, NGC 4589, is shown in Figure 6, where the upperpanel displays the background-subtracted spectra of the central 15′′ region withthe slit positions overlaid on the optical image. The lower panel of Figure 6shows the distribution of the PAH 11.3 μm emission in color as well as that ofthe 5.5 − 6.5 μm continuum emission in contours. The 5.5 − 6.5 μm continuumemission shows a smooth stellar distribution, which is consistent with the 2MASSimage in the upper panel of Figure 6. However, the PAH 11.3 μm emission exhibitsa distinctly elongated distribution, posessing an excellent spatial coincidence withthe minor-axis optical dust lane (see Fig. 2 of Mollenhoff & Bender 1989). Kanedaet al. (2010b) also showed that the molecular hydrogen line emissions come fromthe dust lane, and that the [NeII] line emission shows a more compact distributionnear the nucleus than the PAH features. Thus the PAH emission comes predomi-nantly from dense gas in the optical dust lane, where the properties of the PAHsare not significantly regulated by hard radiation from the LINER nucleus. Since


Fig. 6. Upper panel: the background-subtracted IRS SL (blue) and LL (red) spectra of

the central 15′′ region of NGC4589. The inset shows the SL (left) and LL (right) slit

positions for the spectral mapping, superposed on the 2MASS image. Lower panel: the

distribution of the PAH 11.3 μm emission in color as well as that of the 5.5 − 6.5 μm

continuum emission in contours. For details, see Kaneda et al. (2010b).

NGC 4589 is thought to be an about 108 year old merger (Kawabata et al. 2010),the result may reflect that the PAHs are secondary products from the evolutionof the dust brought in by a merger.

As for the signature of interaction of PAHs with hot plasma, no correlationhas been found between PAH and X-ray emission among elliptical galaxies, whichimplies that a large fraction of observed PAHs are not continuously supplied intothe interstellar space from old stars, but sporadically from other internal or ex-ternal sources. A possible major origin of the PAHs is merger events as suggestedby the above case. Another interesting possibility was suggested by Temi et al.(2007) such that the AGN-assisted feedback outflow from a central dust reservoirmay supply PAHs into the interstellar space of elliptical galaxies through residualdust fragments diminished by sputtering.


5 Summary and future prospects

We extensively review observational results of PAH emission in harsh interstellarenvironments. The observations have indicated that PAHs in the shock regionsof SNRs are destroyed very easily on 10–100 year timescales (e.g. Tycho SNR),where dust grains survive for a much longer time. There are probably a few ex-ceptional cases (3C 396 and N132D in LMC) where SNR still show PAH emission.In contrast, PAHs survive for 106–107 timescales under very intense UV radiationfield in RCW49, one of the most massive star clusters in our Galaxy. The observa-tions have also shown that PAHs are present in the ionized galactic superwinds ofstarburst galaxies. For M 82, PAHs are widely spread around the halo through thegalactic superwind and also probably through past tidal interaction with M 81.For NGC 1569, even very small PAHs emitting the 3.3 μm feature exist in theprominent Hα superwind. Finally the observations have revealed that PAHs arepresent in many elliptical galaxies. The PAH emission from elliptical galaxies ischaracterized by unusually low ratios of the 7.7 μm to the 11.3 μm and 17 μmemission features, probably reflecting that neutral PAHs are dominant over ionizedPAHs due to very soft radiation field from old stars. In particular, for NGC 4589,the PAH emission is found to predominantly come from the optical dust lane. Nosignatures of interaction between PAHs and hot plasmas have been found amongelliptical galaxies, which implies that a large fraction of observed PAHs are spo-radically supplied into the interstellar space from internal or external sources otherthan from old stars.

In future, SPICA (Space Infrared Telescope for Cosmology and Astrophysics;Swinyard et al. 2009) will be launched around 2018, covering the wavelength rangefrom 5 to 200 μm by imaging spectroscopy with a 3-m aperture, 6 K telescope.Detailed studies of PAH emission in the far-IR regime, such as the ro-vibrationalmodes (see Joblin et al. in this volume), can be carried out with SPICA, whichare critical to trace the structural changes of PAH molecules being processed inharsh environments.

Most of the researches presented in this paper are based on observations made with the SpitzerSpace Telescope, which is operated by the Jet Propulsion Laboratory, Carlifornia Institute ofTechnology under NASA contract 1407, and AKARI, a JAXA project with the participation ofESA.

References

Allain, T., Leach, S., & Sedlmayr, E., 1996, A&A, 305, 602


Bendo, G.J., Draine, B.T., Engelbracht, C.W., et al., 2008, MNRAS, 389, 629

Beirao, P., Brandl, B.R., Appleton, P.N., et al., 2008, ApJ, 676, 304

Boulanger, F., Beichman, C., Desert, F.X., et al., 1988, ApJ, 332, 328

Desert, F.X., Boulanger, F., & Puget, J.L., 1990, A&A, 237, 215

Draine, B.T., & Li, A., 2007, ApJ, 657, 810


Engelbracht, C.W., Kundurthy, P., Gordon, K.D., et al., 2006, ApJ, 642, L12

Fukui, Y., Yamamoto, H., Fujishita, M., et al., 2006, Science, 314, 106

Furukawa, N., Dawson, J.R., Ohama, A., et al., 2009, ApJ, 696, 115


Ho, L.C., Filippenko, A.V., & Sargent, W.L.W., 1997, ApJ, 487, 579

Irwin, J.A., Kennedy, H., Parkin, T., & Madden, S.C., 2007, A&A, 474, 461

Irwin, J.A., & Madden, S.C., 2006, A&A, 445, 123

Ishihara, D., Kaneda, H., Furuzawa, A., et al., 2010, A&A, 521, L61

Jones, A.P., Tielens, A.G.G.M., & Hollenbach, D.J., 1996, ApJ, 469, 740

Kaneda, H., Onaka, T., Sakon, I., et al., 2008, ApJ, 684, 270

Kaneda, H., Ishihara, D., Suzuki, T., et al., 2010a, A&A, 514, A14

Kaneda, H., Onaka, T., Sakon, I., et al., 2010b, ApJ, 716, L161

Kawabata, K.S., Maeda, K., Nomoto, K., et al., 2010, Nature, 465, 326

Leipski, C., Antonucci, R., Ogle, P., & Whysong, D., 2009, ApJ, 701, 891

Machida, M., Matsumoto, R., Nozawa, S., et al., 2009, PASJ, 61, 411

Micellota, E.R., Jones, A.P., & Tielens, A.G.G.M., 2010a, A&A, 510, A36

Micellota, E.R., Jones, A.P., & Tielens, A.G.G.M., 2010b, A&A, 510, A37

Mollenhoff, C., & Bender, R., 1989, A&A, 214, 61

Neufeld, D.A., Hollenbach, D.J., Kaufman, M.J., et al., 2007, ApJ, 664, 890

Onaka, T., Matsuhara, H., Wada, T., et al., 2007, PASJ, 59, S401

Onaka, T., Matsumoto, H., Sakon, I., & Kaneda, H., 2010, A&A, 514, A15

Peeters, E., Mattioda, A.L., Hudgins, D.M., & Allamandola, L.J., 2004, ApJ, 617, L65

Reach, W.T., Rho, J., Tappe, A., et al., 2006, AJ, 131, 1479

Reynoso, E.M., Moffett, D.A., Goss, W.M., et al., 1997, ApJ, 491, 816

Shopbell, P.L., & Bland-Hawthorn, J., 1998, ApJ, 493, 129


Sturm, E., Rupke, D., Contursi, A., et al., 2006, ApJ, 653, L13

Swinyard, B. Nakagawa, T., Merken, P., et al., 2009, Experimental Astronomy, 23, 193

Tacconi-Garman, L.E., Sturm, E., Lehnert, M., et al., 2005, A&A, 432, 91

Tappe, A., Rho, J., & Reach, W.T., 2006, ApJ, 653, 267

Telesco, C.M., & Harper, D.A., 1980, ApJ, 235, 392

Temi, P., Brighenti, F., & Mathews, W.G., 2007, ApJ, 666, 222


Van Kerckhoven, C., Hony, S., Peeters, E., et al., 2000, A&A, 357, 1013

Veilleux, S., Rupke, D.S.N., & Swaters, R., 2009, ApJ, 700, L149

Walter, F., Weiss, A., & Scoville, N., 2002, ApJ, 580, L21

Warren, J.S., Hughes, J.P., Badenes, C., et al., 2005, ApJ, 634, 376

Whaley, C.H., Irwin, J.A., Madden, S.C., Galliano, F., & Bendo, G.J., 2009, MNRAS,395, 97

Xilouris, E.M., Madden, S.C., Galliano, F., Vigroux, L., & Sauvage, M., 2004, A&A,416, 41


PAH AND DUST PROCESSING IN SUPERNOVA REMNANTS

J. Rho1, M. Andersen2, A. Tappe3, W.T. Reach1 , J.P. Bernard4

and J. Hewitt5

Abstract. I present observations of shock-processed PAHs and dust insupernova remnants (SNRs). Supernova shocks are one of the primarysites destroying, fragmenting and altering interstellar PAHs and dust.Studies of PAHs through supernova shocks had been limited becauseof confusion with PAHs in background emission. Spitzer observationswith high sensitivity and resolution allow us to separate PAHs asso-ciated with the SNRs and unrelated, Galactic PAHs. In the youngSNR N132D, PAH features are detected with a higher PAH ratio of15–20/7.7 μm than those of other astronomical objects, and we suggestlarge PAHs have survived behind the shock. We present the spectra ofadditional 14 SNRs observed with Spitzer IRS and MIPS SED coveringthe range of 5–90 μm. Bright PAH features from 6.2 to 15–20 μm aredetected from many of SNRs which emit molecular hydrogen lines, in-dicating that both large and small PAHs survive in low velocity shocks.We observe a strong correlation between PAH detection and carbona-ceous small grains, while a few SNRs with dominant silicate dust lackPAH features. We characterize PAHs depending on the shock velocity,preshock density and temperature of hot gas, and discuss PAH anddust processing in shocks and implication of PAH and dust cycles inISM.

1 SOFIA Science Mission Operations/USRA, NASA Ames Research Center, MS 211-3,Moffett Field, CA 94035, USA; e-mail: [email protected] Research & Scientific Support Department, European Space Agency, ESTEC, Keplerlaan1, 2200 AG Noordwijk, The Netherlands3 Harvard-Smithsonian Center for Astrophysics, MS 83, 60 Garden Street, Cambridge,MA 02138, USA4 Centre d’Etude Spatiale des Rayonnements, CNRS, 9 Av. du Colonel Roche, BP. 4346,31028 Toulouse, France5 NASA/Goddard Space Flight Center, Code 662, Greenbelt, MD 20771, USA



1 Introduction

Supernova shocks are one of the primary sites destroying, fragmenting and alter-ing interstellar polycyclic aromatic hydrocarbons (PAHs) and dust. Supernovaeinfluence the chemistry and physics in the interstellar medium (ISM) from galac-tic scales down to the atomic level. Studies of PAHs through supernova shockshad been limited because of confusion with PAHs in background emission. Spitzerobservations with high sensitivity and resolution allow us to separate PAHs asso-ciated with the supernova remnants (SNRs) and unrelated, Galactic PAHs. More-over, ISO observations of the continuum were limited to wavelengths longer than40 μm and the dust analysis was focused on the big grains. Shorter wavelengthobservations are necessary to probe the emission from the very small grains andpolycyclic aromatic hydrocarbons (PAHs) in greater details. The ratio of PAHand very small grain (VSG) to big grain (BG) abundance can provide insight ofthe dust destruction mechanisms.

Dust grains and PAHs control the heating and cooling balance of interstellargas through absorption of ultraviolet radiation, generation of photoelectrons, andemission of infrared radiation. These processes allow us to use them as diagnosticsto probe physical conditions in a range of diverse environments. However, thisis hampered by our lack of knowledge of their exact structure and composition,and how these are in turn affected by the physical environment. The PAHs arebelieved to give rise to the unidentified infrared bands (UIR), well-known emissionfeatures near 3.3, 6.2, 7.7, 8.6, and 11.2 μm attributed to the CC and CH stretchingand bending modes of these molecules (Allamandola et al. 1989). More recentlyobserved features at 15–20 μm are interpreted as PAH CCC out-of-plane bendingmodes (Peeters et al. 2004). Several detections have been reported in a variety ofastronomical objects.

We present observations of shock-processed PAHs and dust in SNRs withSpitzer IRAC and MIPS imaging and IRS and MIPS-SED spectroscopy coveringfrom 3.6 μm to 95 μm. We confirm the PAH detection and its distribution in theyoung SNR N132D and present GLIMPSE-selected (Reach et al. 2006) 14 SNRs,many of which are interacting molecular clouds. We characterize PAHs depend-ing on a shock velocity, preshock density and temperature of hot gas, and discusscorrelation of PAH and associated dust composition and dust cycles in ISM.

2 PAH detection from the supernova remnant N132D

N132D is a young SNR in the Large Magellanic Cloud (LMC). The age of N132Dis 2500 yr. In the optical, the large-scale morphology of the gas in N132D showshighly red-shifted oxygen-rich ejecta in the center and shocked interstellar cloudsin the outer rim. We have previously presented Spitzer IRAC, MIPS and anIRS spectrum of N132D (Tappe et al. 2006). Here we present IRS-LL mappingof N132D covering the entire SNR, and IRS staring of SL observations toward4 positions.

J. Rho et al.: PAH and Dust Processing in Supernova Remnants 171

Fig. 1. N132D spectral mapping (center; blue = [Ne III] 15.5 μm, green = [Fe II] 17.9 μm,

red = MIPS 24 μm, white dashed line = red and blueshifted high-velocity Ne/O ejecta

seen in the optical with HST) and spectra of selected positions.

We detect bright 24 and 70 μm emission from shocked, heated dust grains withMIPS. The 5–40 μm IRS mapping was performed, where LL (13–40 μm) coveredthe entire SNR and SL (5–13 μm) covered a few interesting positions. Figure 1shows representative sets of spectra from southern rim (bottom left spectrum), theWest Complex (bottom right), Lasker Bowl (top right) and an ejecta knot (topleft).

IRS spectrum toward the southeastern shell shows a steeply rising continuumwith [Ne III] and [O IV] as well as PAH features at ∼7.7, 11.3, and 15–20 μm. ThePAH bands are from swept-up ISM material, which are processed by the strongshock wave via thermal sputtering and grain-grain collisions (Tielens et al. 1994).We found evidence for spatial variations of the PAH features, both in intensityand wavelength, when comparing spectra extracted along the IRS long-low slitfrom the shell of the remnant and the adjacent molecular cloud/H II region, whichconfirms the processing of the PAH molecules. These detections, most notablythe 15 ∼ 20 μm features, are the first of its kind for supernova remnants toour knowledge (Tappe et al. 2006). The 15 ∼ 20 μm features in Figure 2a areinterpreted as PAH CCC out-of-plane bending modes (cf. van Kerckhoven et al.2000; Peeters et al. 2004). We generated a PAH map between 15–20 μm (Fig. 2b)after subtracting the continuum at neighboring wavelengths; the morphology issimilar to that of the dust continuum in the MIPS 24 μm band (Fig. 1). Severaldetections of 15–20 μm PAH features have been reported in a variety of sources,including H II regions, young stellar objects, reflection nebulae, and evolved stars(see Peeters et al. 2004, for references therein). The detections include both broadvariable plateaus as well as distinct, narrow features, most notably at 16.4 and17.4 μm.


Fig. 2. (a: Left) N132D southeastern rim spectrum (magenta line = SL2, red line = SL1,

green line = LL2, blue line = LL1): the local background and the modified blackbody

continuum have been subtracted. The total error includes the reduction pipeline delivered

errors and an estimate of the variation in the local background (Tappe et al. 2006). (b:

Right) PAH grey-scale map of N132D between 15–20 μm after subtracting neighboring

wavelength dust continuum, superposed on MIPS 24 μm contours from Tappe et al.

(2006).

3 PAH from slow shocks: Molecular supernova remnants

Interacting SNRs are an ideal laboratory to study the effects of fast shocks onthe interstellar material. Shocks can both sputter and shatter interstellar dustand thus potentially modify its abundance and size distribution. Previous obser-vations of interacting SNRs had been limited. Using Spitzer GLIMPSE surveydata (Reach et al. 2006) identified a series of Galactic SNRs in the infrared, a sub-sample having IRAC colors that indicate shocked H2. We present follow–up SpitzerIRS low resolution and MIPS SED spectra of a sample of 14 SNRs – G11.2-0.3,G21.8-0.6 (Kes 69), G22.7-0.2, G39.2-0.3 (3C 396), G54.4-0.3, G304.6+0.1 (Kes17), Kes 20A (G310.8-0.4), G311.5-0.3, G332.4-0.4 (RCW 103), G344.7-0.1, CTB37A (G348.5+0.1), G348.5-0.0, and G349.7+0.2. Molecular hydrogen detectionstowards 12 SNRs are presented by Hewitt et al. (2009) and Andersen et al. (2010),more than doubling the sample of known interacting Galactic SNRs. Here wepresent an analysis of the continuum and PAH emission. The spectroscopy ob-servations were centered on the emission peaks for each SNR. The wavelengthcoverage from 5 to 90 μm ensures a good sampling of the three main dust speciesand enables us to fit the continuum in greater detail.

The dust model fitting of the SEDs (Fig. 3a) is done using the method describedin Bernard et al. (2008). Three dust components are adopted, PAH molecules, VerySmall Grains (VSG) and Big Grains (BG). The abundance of each dust species(YPAH , YV SG, YBG) and the strength of the radiation field (XISRF ) are taken


Fig. 3. (a: Left) the DUSTEM fit for 3C 396 superposed IRS and MIPS SED data with

3 component models of PAH, very small grains and big grains. The PAH, very small

and large grain contribution are shown as the dotted line, the thin solid line, and the

dashed dotted. The total fit is shown as the thick solid line (from Andersen et al. 2010).

(b: Right) comparison of PAH emission in the SNR Kes 17 (red curve), Kes 69 (dark

blue) and N132D (orange) with those of IRAS11308 (green), LkHα 234 (PDR: blue) and

NGC7331 (yellow).

as free parameters in the fit. For all three dust species, a MRN powerlaw sizedistribution is assumed. The slopes and size ranges are different than adopted inthe original version of DUSTEM due to the updated input physics, mainly thecross sections adopted have changed (Compiegne et al. 2008). The reduced χ2

range between 4–10. Systematically higher reduced χ2 around the 15–20 μm showthat the PAH features in SNRs are significantly different from those in PDRs. Thefit results are summarized in Figure 4a, showing that abundances of PAHs andVSGs are higher than those in the Milky Way and the LMC (Bernard et al. 2008).

A dust model composed of PAHs, VSGs and BGs provide a good fit to the SEDsof the SNRs. All the SNRs show evidence for PAH emission. Two SNRs, G11.2-0.3, and G344.7-0.1 show no to little evidence for VSGs and a lower PAH/BG ratiothan observed in the diffuse interstellar medium and the LMC. We show Kes 17and Kes 69 as examples (Fig. 3b) to demonstrate the differences in PAH emissionbetween dust processed by a SNR shock and PDR dominated dust. Typically theradiation field is 10–100 times larger than normal interstellar radiation ISRF, thestrength of which is consistent with being created from the shock.

The dust spectral fitting indicates the presence of PAH emission in most ofSNRs. In Figure 3b, PAH spectra of Kes 17 and Kes 69 are compared with the


Fig. 4. (a: Left) the ratios of PAHs and very small grains to big grains are higher than

those of the Milky Way (solid line). Two young SNRs (G11.2 and G344.2) have the ratios

of carbon dust to silicates smaller than the Milky Way value and are associated with a

strong radiation field. (b: Right) IRS spectrum of the Puppis A SNR superposed on the

best fit (solid line) of the ZDA model (Arendt et al. 2010). The dotted line is a model

with PAH and VSG, which shows similarity to the N132D spectrum.

emission from LkHα 234 and NGC7331, a PDR and a nearby spiral galaxy withstrong PAH emission, respectively. We find that the PAH features of molecularinteracting (H2 emitting) SNRs are very similar to each other as shown for thecases of Kes 17 and Kes 69. The PAH emission is detected in our backgroundspectra since all the SNRs are located in the Galactic plane. However, the PAHemission is clearly enhanced in the shocked region and is present after backgroundsubtraction. The 15–20μm PAH feature in Kes 17 is stronger than that of LkHα234 whereas the PAH features between 5–14 μm are almost identical. It is similarto that of NGC 7331. The ratio of the emission from the 15–20 μm plateau tothe 6.2 μm PAH feature is ∼0.4 for Kes 17 and other H2 emitting SNRs, similarto NGC 7331 but larger than for the PDR. This ratio is smaller than that of theyoung SNR N132D (Tappe et al. 2006). The difference in PAH features betweenmolecular SNRs and N132D is due to the environments of SNRs. Many of theinteracting SNRs are in a dense molecular cloud environment with a slow shockwhile N132D is less dense environment and has a strong shock which significantlydestroy small PAHs.


3.1 Comparison with Puppis A and other dust model

Figure 3b shows IRS spectrum of a middle-age SNR, Puppis A. Spectral fittingof Puppis A (Arendt et al. 2010) was performed based on carbon and silicatedgrains using Zubko et al. (2004; hereafter ZDA), radiatively heated by an ambientinterstellar radiation field. The ZDA models are a set of models for the dust grainsize (or mass) distributions that are derived to simultaneously fit average interstel-lar extinction, emission, and abundances. Different models are distinguished bydifferent choices of grain compositions and different sets of abundance constraints.The PAH emission is not evident in Puppis A. In contrast, PAH features are no-ticeable in N132D, implying that the grains in N132D is less sputtered than thosein Puppis A.

References

Andersen, M., Rho, J., Tappe, A., et al., 2010, submitted

Arendt, R.G., Dwek, E., Blair, W.P., et al., 2010, ApJ, 725, 585

Bernard, J.P., Reach, W.T., et al., 2008, AJ, 136, 919

Compiegne, M., Abergel, A., Verstraete, L., & Habart, E., 2008, A&A, 491, 797

Hewitt, J.W., Rho, J., ndersen, M., & Reach, W.T., 2009, ApJ, 649, 1266

Peeters, E., Mattioda, A.L., Hudgins, D.M., & Allamandola, L.J., 2004, ApJ, 617, L65

Reach, W.T., et al., 2006, AJ, 131, 1479

Tappe, A., Rho, J., & Reach, W.T., 2006, ApJ, 653, 267

van Kerckhoven, C., et al., 2000, A&A, 357, 1013

Zubko, V., Dwek, E. & Arendt, R.G., 2004, ApJS, 152, 211


THE FORMATION OF POLYCYCLIC AROMATICHYDROCARBONS IN EVOLVED CIRCUMSTELLAR

ENVIRONMENTS

I. Cherchneff1

Abstract. The formation of Polycyclic Aromatic Hydrocarbons in thecircumstellar outflows of evolved stars is reviewed, with an emphasison carbon stars on the Asymptotic Giant Branch. Evidence for PAHspresent in their winds is provided by meteoritic studies and recentobservations of the Unidentified Infrared bands. We detail the chemicalprocesses leading to the closure of the first aromatic ring as well asthe growth mechanisms leading to amorphous carbon grains. Existingstudies on PAH formation in evolved stellar envelopes are reviewed andnew results for the modelling of the inner wind of the archetype carbonstar IRC+10216 are presented. Benzene, C6H6, forms close to the star,as well as water, H2O, as a result of non-equilibrium chemistry inducedby the periodic passage of shocks. The growth process of aromatic ringsmay thus resemble that active in sooting flames due to the presenceof radicals like hydroxyl, OH. Finally, we discuss possible formationprocesses for PAHs and aromatic compounds in the hydrogen-rich RCrB star, V854 Cen, and their implication for the carriers of the RedEmission and the Diffuse Interstellar Bands.

1 Introduction

Upon the suggestion by Leger & Puget (1984) and Allamandola et al. (1985) thatpolycyclic aromatic hydrocarbons (PAHs) were responsible for the emission of aseries of unidentified infrared (UIR) bands ubiquitous to the interstellar medium(ISM), large efforts have been devoted to understand the presence of these aromaticmolecules in space. PAHs are commonly found on Earth in the exhaust fumesof car engines and are key intermediates in the inception and growth of sootparticles in incomplete combustion processes (Haynes & Wagner 1981). They areassociated to the fine particles called aerosols released by human activities in the

1 Departement Physik, Universitat Basel, Klinglebergstrasse 82, 4056 Basel, Switzerland



Earth troposphere. PAHs are also mutagens and thus carcinogens. Therefore,PAHs are extensively studied on Earth, providing an abundant and exhaustivescientific literature of great interest to the study of PAHs in space. Cosmic PAHsare observed through their UIR bands in several local, low- and high-redshiftenvironments including HII regions, Proto- and Planetary Nebulae, young stellarobjetcs, and the ISM of high redshift galaxies (Tielens 2008). We can define twopopulations of PAHs in space: 1) PAHs formed directly from the gas phase indense media; and 2) PAHs as products of dust sputtering in harsh environments.These two distinct classes require very diverse environments and processes to form,the second class being prevalent in galaxies and most readily observed through thedetection of the UIR bands. In this review, we focus on the first class of PAHs,those forming under specific conditions from the gas phase in evolved circumstellarenvironments. We discuss evidence for PAHs in stardust inclusions in meteoritesand PAHs observed in evolved low-mass stars. We detail the formation routes tobenzene and its growth to several aromatic rings. We review existing models andpresent new results on benzene formation in the extreme carbon star IRC+10216.Finally, the PAH synthesis in R CrB stars is discussed.

2 Carbon dust providers in our galaxy

PAHs are intimately linked to the formation of amorphous carbon (AC) dust orsoot from the gas phase in combustion processes on Earth. It is thus natural reck-oning that similar formation conditions may exist in space that are conducive to thesynthesis of PAHs. Those conditions characteristic of the gas phase are: (1) aninitial carbon-, hydrogen-rich chemical composition; (2) high densities; (3) andhigh temperatures. Such conditions are usually found in space in circumstellarenvironments such as the winds or ejecta of evolved stars. Furthermore, theseevolved objects are the prominent providers of dust in our galaxy. Therefore, theformation of PAHs in space from the gas phase is linked to evolved stellar objectssynthesising AC dust. The prevalent AC grain makers in our galaxy include thelate stages of evolution of low- and high-mass stars. They are the wind of low-masscarbon-rich stars on the Asymptotic Giant Branch (AGB), the ejecta of Type IIsupernovae, the colliding winds of carbon-rich Wolf-Rayet stellar binaries, and theclumpy winds of R CrB stars. The physical and chemical conditions pertainingto these environments are summarized in Table 1. For carbon AGB stars, thegas layers close to the photosphere that are periodically shocked satisfy the re-quired conditions for PAH formation (Cherchneff 1998). Conversely, the ejecta ofsupernovae do not as the hydrogen present in the gas is not microscopically mixedwithin the ejecta layers but is in the form of macroscopically-mixed blobs of gas(Kifonidis et al. 2006). Carbon-rich Wolf-Rayet stars have lost their hydrogenenvelopes and their wind is H-free. The observed AC dust is thus synthesised by achemical route that does not involve PAHs but rather bare carbon chains and rings(Cherchneff et al. 2000). However, WC stars are part of binary systems in whichthe companion is usually a OB star characterised by a H-rich wind, thus implyingthat hydrogen is present in the colliding wind region. But owing to the high gas

I. Cherchneff: PAHs in Evolved Circumstellar Environments 179

Table 1. Most important amorphous carbon (AC) dust stellar providers in our galaxy.

Listed in the first column are: (1) the amount of AC dust formed over the star lifetime,

(2) the region where dust forms, (3) the presence of hydrogen, (4) the gas density and

temperature at which dust forms, and (5) the observed molecules related to AC dust

synthesis. Value for supernovae are for SN 1987A (Ercolano et al. 2007) and for the

SN progenitor of Cas A (Nozawa et al. 2010). Value for R CrB stars is an upper limit

assuming all carbon condenses in AC dust.

Star Carbon AGB Type II Supernova

AC mass (M�) 3 × 10−3 – 1 × 10−2 7×10−4 – 7 × 10−2

Dust locus Shocked inner wind EjectaHydrogen Yes Yes – not mixed with CGas density (cm−3) 108–1013 109–1012

Gas temperature (K) Low: 1000 – 1500 High: 3000Key species C2H2 CO & SiO

Star Carbon-rich Wolf-Rayet R CrB

AC mass (M�) 0.1 2 × 10−6

Dust locus Colliding winds Episodic clumpsHydrogen No No except V854 CenGas density (cm−3) 1010 109–1011

Gas temperature (K) High: 3000 Medium: 1500 – 2500Key species No observation yet C2

temperatures, it is unlikely that the chemical routes responsible for PAHs synthe-sis may be active in these hot regions. Finally, R CrB stars are extremely rare,evolved low-mas stars exhibiting supergiant photospheric conditions and episodi-cally ejecting clumps of gas rich in AC dust. Their wind is H-free but there existsone object, V854 Cen, which is H-rich and for which the UIR bands have beenobserved with ISO (Lambert et al. 2001). PAHs may therefore play a role in theformation process of AC dust in this specific object.

Studies on the formation of PAHs in circumstellar environments have focusedon the prevalent AC dust makers, the carbon AGB stars, represented by the well-studied carbon star IRC+10216. We will thus concentrate on these stellar dustproviders in the rest of this review.

3 Evidence for circumstellar PAHs in meteorites

Crucial information on the chemical composition, structure, and formation locusof stardust has been gained since the isolation of presolar grains in primitive me-teorites and their study by laboratory microanalyses. Graphite grains bearing theisotopic composition resulting from the nucleosynthesis at play in low-mass starshave been isolated implying a formation locus in AGB outflows (Zinner 1997).Two basic graphite spherule morphologies are found, designated by “onion” and


cauliflower’ (Bernatowicz et al. 2006). Two thirds of the onion-type spherules arecharacterised by concentric outer shells of well-crystallised graphite surroundinga nanocrystalline carbon core consisting of graphene curled sheets indicative ofpentagonal ring insertion. One quarter of this core mass is in the form of smallPAHs. Small aromatics including phenantrene (C14H10) or chrysene (C18H12)have further been identified in the study of presolar graphite onion spherules.Those aromatics bore similar isotopic anomalies than their parent graphite cir-cumstellar grains, indicating a possible formation in AGB winds (Messenger et al.1998). The cauliflower spherules are formed of contorted graphene sheets with nonanocrytalline core. The morphology of these two types of spherules hint at dustcondensation scenarios in AGB winds. The onion morphology indicates that theinner gas layers where dust forms are hotter, with long enough residence times toform dust precursors like PAHs and graphene sheets and allow the outer layers ofthe newly-formed dust grain to become well graphitised. The cauliflower struc-ture on the other hand is indicative of a lower temperature inner wind and rapidmotion outward of the shocked layers above the photosphere. Dust precurors likePAHs have formed and are included into graphene sheets but not enough time isgiven for these amorphous structures to become crystalised. An amorphous, lessstructured carbon spherule thus results from these milder inner wind conditions.

Therefore, there exists direct evidence of the presence of PAHs in the dustsynthesis zone of carbon-rich AGB stars from the study of presolar grains extractedfrom meteorites. Moreover, it appears that PAHs do form as precursors before theend of the dust condensation process as found in the pyrolysis of hydrocarbonsand in combustion processes like hydrocarbon-fueled flames (Frenklach & Warnatz1987; Richter & Howard 2000).

4 Evidence for circumstellar PAHs from observations

The common fingerprints of PAHs in space are the emission of the UIR bandsthrough excitation by a single ultraviolet (UV) photon. These bands can be de-tected in any circumstellar medium, providing the existence of a UV radiationfield. The photospheric temperatures of carbon AGB stars are too low to pro-duce a strong UV stellar radiation field capable of exciting PAH molecules in thedust formation zone. Buss et al. (1991) and Speck & Barlow (1997) observed thecarbon star TU Tauri, part of a binary system with a blue companion providingfor UV photons, and detected possible UIR bands. More recently, Boersma et al.(2006) analysed the SWS ISO spectra of several carbon stars including TU Tauand found UIR emission bands in the latter only. They ascribed the UIR spectrumto PAH excitation by the UV radiation field of the companion star when other sin-gle carbon stars in the sample were devoid of UIR bands in their spectrum. Theypointed out that PAHs may be present in any carbon star but not observable dueto the lack of exciting radiation. Another probe to the presence of PAHs in AGBstars is the study of S stars, i.e., oxygen-rich AGB stars on the verge of dredgingup carbon in their photosphere to become carbon stars. They are characterisedby a photospheric C/O ratio less than but �1 and by the fact that they do not yet


Fig. 1. Left: growth of PAHs according to the H abstraction-C2H2 addition (HACA)

mechanism proposed by Frenklach et al. (1984). Right: growth of aromatic structures

via the dimerisation and coalescence of PAHs proposed by Mukhergee et al. (1994).

produce carbon dust. Smolders et al. (2010) analysed IR low resolution spectrataken by Spitzer of several galactic S stars. Four of them characterised by C/O ra-tios very close to 1 show the UIR bands in their spectra. Among those, three starsshow weak extended 6.2 μm emission bands characteristic of PAHs with a mixtureof aliphatic and aromatic bonds excited by a weak UV field. S stars possess a dualchemistry in their dust formation zone due to the pulsation-induced shocks asshown by Cherchneff (2006). As a result of non-equilibrium chemistry, O-bearingmolecules form close to the stellar photosphere whereas hydrocarbons (e.g. C2H2)form at larger radii (∼3 R�). PAHs may thus form at lower temperatures anddensities, favouring small disordered aromatic structures linked by aliphatic bonds.At these radii, the formation of AC dust may be hampered by the low densitiesand residence time resulting in a population of small aromatic/aliphatic structuresthat can not turn into dust but are responsible for the observed UIR bands.

5 Chemical routes to PAH formation in flames

Incomplete combustion produces soot particles where PAHs are observed to playa key role in the inception and growth process of AC grains (Richter & Howard2002). The formation of the single aromatic ring benzene (C6H6) represents therate limiting step to soot synthesis as it is the natural passage from the aliphaticstructure of the fuel hydrocarbons to the closed ring aromatic structures whichare the backbones of combustion products like soot. Three dominant routes tofirst ring closure have been identified in flames. The prevalent route involves therecombination of two propargyl radicals (C3H3) to form cyclic and linear benzeneand the benzene radical phenyl (C6H5) according to the following reactions

C3H3 + C3H3 −→ C6H6 (5.1)


andC3H3 + C3H3 −→ C6H5 + H. (5.2)

These routes were first studied and proposed by Miller & Mellius (1992) as im-portant sources of aromatic rings in flames. The propargyl recombination intoone aromatic ring involves many reactive channels leading to the formation ofseveral intermediates such as fulvene (C5H4CH2), observed in flames (Miller &Klippenstein 2003; Tang et al. 2006). In pyrolytic shock tube study of propargylreactions, Scherer et al. (2000) show that the direct formation of benzene viaReaction 5.1 is prevalent over 5.2.

The other two chemical pathways to first aromatic ring closure in acetylenicflames involve the reaction of 1-buten-3-ynyl, 1-C4H3 with acetylene proposed byFrenklach et al. (1984), and the reaction of 1,3-butadienyl, 1-C4H5 with acetyleneproposed by Cole et al. (1984). They lead to the following final products via directpathways involving chemically activated isomerizations (Westmorland et al. 1989)

1 − C4H3 + C2H2 −→ C6H5 (5.3)

and1 − C4H5 + C2H2 −→ C6H6 + H. (5.4)

In the context of PAH formation in circumstellar outflows, the above pathwayswill occur if the reactants are present in the gas. The formation of C3H3 resultsfrom the reaction of C2H2 and methylene, CH2, whereas vynyl-acetylene, C4H4

will form C4H3 and C4H5 from its reaction with atomic hydrogen.Once benzene is made available in the gas, larger PAHs may grow according

to three different scenarios. The first scenario was proposed by Frenklach et al.(1984) and involves the growth of large aromatic structures through sequentialhydrogen abstraction and acetylene addition (referred as the HACA mechanism).The formation of radical sites on the aromatic structure enables C2H2 to add andfinally form a new aromatic ring as depicted in Figure 1. The second route wasproposed by Mukherjee et al. (1994) in their study of the pyrolysis of pyrene(C16H10). In the absence of PAH species arising from the HACA growth pathway,they derive that the dominant weight growth channel was the direct polymerisationof PAHs through the initial formation of PAH dimers, as illustrated in Figure 1 forpyrene. Finally, a third growth pathway was proposed by Krestinin et al. (2000)in their study of acetylene pyrolysis and involves the polymerization of polyyneson a surface radical site of any small grain seed as illustrated in Figure 2. Thishypothesis is supported by the observation of small polyynes (C4H2, C6H2, C8H2)along with PAHs in the sooting zone of flames (Cole et al. 1984).

One should bear in mind that there still exists controversy on the prevalentpathways to AC grain growth (for more detail, see Jager et al., this volume).Despite the fact that the HACA mechanism is largely used to describe the growthof aromatic rings, it does not provide a satisfactory explanation for the earlyformation in sooting flames of small carbonaceous nanoparticles made of a fewaromatic rings linked by aliphatic bonds (Minutolo et al. 1998).


Fig. 2. Polymerization of polyynes on a surface radical site and its transformation into

aromatic structures leading to breeding of radical sites and rapid growth (from Krestinin

2000).

6 PAH yields from existing studies

Only a few studies have been conducted to date on the formation of PAHs inthe outflows of carbon AGB stars. Frenklach & Feigelson (1989) first applied theHACA mechanism to the steady wind of carbon stars to derive dust masses andPAH yields. They highlighted the existence of a temperature window (900 K –1100 K) for which PAH growth occurred. However, their treatment of the stellarwind was not taking into account the action of shocks active above the stellarphotosphere. Therefore, the gas density dropped by orders of magnitude over shorttime scales impeding the efficient synthesis of PAHs and dust grains. Cherchneffet al. (1992) studied the formation of PAHs using the HACA mechanism forPAH growth but including the newly proposed Reactions 5.1 and 5.2 for first ringclosure. They include a treatment of the effects of shocks on the gas above thephotosphere but also used a Eulerian treatment for the steady wind. The resultingPAH yields were small for all the various wind models and the shocked layersclose to the stars experiencing quasi-ballistic trajectories were proposed as possibleloci for a more efficient synthesis of PAHs. Cadwell et al. (1994) studied theinduced nucleation and growth of carbon dust on pre-existing seed nuclei. FinallyCherchneff & Cau (1992) and Cau (2002) investigated the formation of PAHsand their dimers in the shocked inner wind of IRC+10216, using a Lagrangianformalism that follows the post-shock gas trajectories induced by the periodicshocks and the stellar gravitational field, following a study from Bowen (1988). Theformation of PAHs and their dimers occurred at higher temperatures than thoseof Frenklach & Feigelson, typically for T ≤ 1700 K, in agreement with a recentlaboratory study of soot formation in low-temperature laser-induced pyrolysis ofhydrocarbons by Jager et al. (2009). The deduced PAH yields were sufficient toaccount for the mass of PAHs included in the nanocrystalline cores of graphitespherules. However, they could not account for the total mass of AC dust formedin the wind, indicating that further dust growth mechanisms must occur such asthe deposition of acetylene on the surface of dust seeds.

In view of these existing theoretical studies and new observational results,the ubiquity of PAHs in the inner wind of carbon AGB stars is almost certain.


Fig. 3. A schematic representation of the inner envelope of IRC+10216. The gas param-

eters span the values indicated as a function of radius. The quasi-balastic trajectories

of the gas layers are shown according to Bowen (1988). Dust grains are represented by

small squares and balls. The mentioned molecules form according to TE and non-TE

chemistry depending on their locus. PAHs form in the inner wind between 1 R� and

∼5 R�.

However, a complete model that includes the various pathways of PAH formationand growth to AC dust as described in Section 5, and applied to the shocked innerlayers of carbon stars is still lacking.

7 New results on benzene formation in IRC+10216

The carbon star IRC+10216 has been extensively studied due to its proximityand its stage of evolution that guarantees an extremely rich chemistry in its wind(Olofsson 2006). The dust forms close to the star in gas layers that experience theperiodic passage of shocks induced by the stellar pulsation. The energy spectraldistribution of IRC+10216 is well reproduced by a dust composition mainly dom-inated by amorphous carbon with a small contribution from silicon carbide andmagnesium sulphide (Keady et al. 1988; Ivezic & Elitzur 1996; Hony et al. 2002).Acetylene has been detected in the mid-infrared close to the star with high abun-dances (Keady & Ridgway 1993; Fonfrıa et al. 2008). A schematic representationof the inner wind of IRC+10216 is given in Figure 3 with the gas temperatureand density as a function of radius, the various chemical states of the gas and


Fig. 4. Molecular abundances with respect to total gas as a function of pulsation period

phase at 1.5 R�. The shock strength at this radius is 17.9 kms−1.

the molecules observed. Briefly, molecules are formed at chemical equilibrium inthe photosphere owing to the high gas temperatures and densities. They are thenreprocessed by the periodic passage of shocks induced by the stellar pulsation. Inthe hot postshock gas, molecules are destroyed by collisions and reform during theadiabatic cooling of the gas as shown in Figure 3. Prevalent species (e.g. CO,HCN, CS, C2H2) reform with abundances close to their TE values whereas thecollisional breaking of CO after the shock front induces a new chemistry responsi-ble for the formation of ’exotic’ molecules of lower abundances and not expectedto form in C-rich environments. As an example, the radical hydroxyl, OH, quicklyforms and triggers the synthesis of O-bearing species like SiO in the inner wind(Willacy & Cherchneff 1998). Thus, an active non-equilibrium chemistry is re-sponsible for the reprocessing of molecules in the dust formation zone of AGBstars (Cherchneff 2006).

PAHs do not escape this shock molecular reprocessing and form once the adia-batic cooling of the shocked gas layers produces the temperature range conduciveto the closure of the first aromatic ring (Cherchneff & Cau 1998; Cau 2002). Wemodel the inner wind of IRC+10216 using the Lagrangian formalism of Willacy &Cherchneff (1998) but considering gas densities lower by one order of magnitudethan their values for those were slightly overestimated (Agundez & Cernicharo2006). An updated chemistry is used that includes the formation of the ma-jor O/Si/S/C-bearing molecules as well as the routes to the closure of the firstaromatic ring described in Section 5. The results for the adiabatically-cooledpost-shock gas layer located at 1.5 R� are shown in Figure 4, where molecularabundances with respect to total gas are presented versus phase of the pulsation


period. As previously mentioned, benzene, C6H6 forms once the gas has cooledto temperatures <1700 K, while phenyl, C6H5 is consumed in benzene throughits reaction with H2. The formation of C6H5 is triggered by the recombination ofpropargyl radicals, C3H3, according to Reaction 5.2. The low abundances of C4H2

hint at the fact that PAH growth beyond benzene cannot proceed through polyynepolymerization but may rather involve the HACA mechanism and/or the directcoalescence of PAHs. As expected, the abundances of small carbon chains C2 andC3 peak at very early phases where the high temperatures precludes the formationof aromatics. However, their abundances are always low for the overabundance ofhydrogen favours the formation of hydrocarbons over bare carbon chains.

A secondary important result is the formation of water, H2O, in the innerwind of IRC+10216. Our results show that H2O already forms at radius 1.5 R�.OH is quickly synthesised after the shock front at early phases by the reactionof atomic oxygen stemming from the partial dissociation of CO with the newlyrecombined H2. H2O is then synthesised from the reaction of OH with H2. Wa-ter has recently been detected with Herschel/HIFI in the carbon star V Cygni(Neufeld et al. 2010) and the S star Chi Cygni (Justtanont et al. 2010), andwith Herschel/SPIRE/PACS in IRC+10216 (Decin et al. 2010). Those detectionsimply an inner origin for H2O. Several mechanisms have been proposed since thefirst detection of H2O in IRC+10216 (Melnick et al. 2001), and include the vapor-isation of icy comets orbiting carbon stars (Melnick et al. 2001), the synthesis ofwater on dust grain surfaces in the intermediate envelope (Willacy 2004), and thephotodissociation of 13CO by interstellar UV photons penetrating clumps (Decinet al. 2010). Our preliminary results point to a more genuine mechanism for waterformation which results from the non-equilibrium chemistry induced by shocks inthe inner wind of IRC+10216. The presence of water in the dust formation zoneindicates that PAH formation and growth processes may be close to those found inacetylenic flames where oxidation by atomic O and O2 occurs. While the amountsof atomic O and dioxygen remain low at all phases in the gas layers under study,OH may attack propargyl radicals to form the formyl radical HCO. Therefore, thebenzene abundances and yields shown in Figure 3 may be overestimated due to thefact that the oxidation of hydrocarbons has not yet been included in the chemistry.Nevertheless, our preliminary results indicate a rapid conversion of hydrocarbonsinto monoaromatic rings.

8 Other environments conducive to PAH synthesis

Apart from the winds of carbon AGB stars, there may exist other circumstellarmedia where PAHs form. One such environment is the wind of one hydrogen-rich R Corborealis star, V854 Centauri. R CrB stars are extremely rare, evolved,low-mass stars showing supergiant photospheric signatures. They have lost theirentire hydrogen shell and experience strong fading on periods of the order of afew years. This sudden drop in luminosity is ascribed to the random ejection ofAC dust clouds along the line of sight. V854 Cen is a peculiar R CrB as it hasretained some hydrogen in its wind. Of particular interest is the detection of red


emission (RE) bands at minimum whose wavelengths match those of a series ofemission bands observed in the proto-planetary nebula, the Red Rectangle (Rao &Lambert 1993a). Furthermore, those frequencies almost match some of the DiffuseInterstellar Bands (DIBs) observed in the ISM. The UIR bands were also detectedwith ISO in V854 Cen pointing to a PAH component in the stellar wind while theSwan and Phillip bands of C2 were observed during the decline to light minimum(Rao & Lambert 1993b, 2000; Lambert et al. 2001). The stellar temperature ishigh (∼7500 K) indicating that despite the presence of hydrogen, the formationof bare carbon chains should proceed, as confirmed by the presence of C2. As thetemperature of the cloud drops with expansion and because of the surroundinghydrogen, polyyne molecules could form along with already formed carbon mono-cyclic and aromatic rings, leading to a growth mechanism for dust involving thepolymerization of polyynes on radical sites (Krestinin et al. 2000). This growthmechanism should partly result in a left-over population of nanoparticles made ofunits of one or two aromatic rings linked by aliphatic bonds, as observed at hightemperatures in sooting flames (Minutolo et al. 1998). Hence the formation ofPAHs in V854 Cen occurs along chemical pathways different from those involved inAGB winds (e.g., the HACA mechanism) and these aromatic nanoparticles couldbe serious contenders as carriers of the RE and the DIBs.

9 Conclusion

The formation of PAHs from the gas phase and their role in the synthesis of carbondust occurring in hydrogen-rich, low-temperature circumstellar environments likeAGB stellar winds are well accepted. Small PAHs act as building blocks in theformation of carbonaceous nanoparticles precursors to dust seeds whereas largeones are direct intermediates to the synthesis of AC grains. A population of smallPAHs will survive the AC dust condensation zone and is probably responsiblefor the UIR bands observed in certain AGB winds. However, most PAHs formedin the inner gas layers will be quickly included into AC dust grains and in thiscontext, carbon AGB stars are not important PAH providers to the ISM. Despitethe fact that circumstellar PAHs are less studied than their interstellar ubiquitouscounterparts stemming from AC dust processing in the ISM, they are as important,being at the origin of the carbon dust formation process in evolved stars.

References

Agundez, M., & Cernicharo, J., 2006, ApJ, 650, 374


Bernatowicz, T., Croat T.K., & Daulton, T.L., 2006, M.E.S.S., 943, 109

Boersma, C., Hony, S., & Tielens, AGGM, 2006, A&A, 447, 213

Bowen, G.H., 1988, ApJ, 329, 299

Buss, R.H. Jr, Tielens, A.G.G.M., & Snow, T., 1991, ApJ, 372, 281

Cadwell, B.J., Wang, H., Feigelson, E.D., & Frenklach, M., 1994, ApJ, 429, 285


Cau, P., 2002, A&A, 392, 203

Cherchneff, I., Barker, J.R., & Tielens, A.G.G.M., 1992, ApJ, 401, 269

Cherchneff, I., 1998, ISAA, 4, 265

Cherchneff, I., & Cau, P., 1999, in Asymptotic Giant Branch Stars, IAU Symp. 191, ed.T. Le Bertre, A. Lebre, & C. Waelkens, 251

Cherchneff, I., LeTeuff, Y., Williams, P.M., & Tielens, A.G.G.M., 2000, A&A, 357, 572

Cherchneff, I., 2006, A&A, 456, 1001

Cole, J.A., Bittner, J.D., Longwell, J.B., & Howard, J.B., 1984, Combustion and Flame,56, 51

Decin, L., Agundez, M., Barlow, M.J., et al., 2010, Nature, 467, 64

Ercolano, B., Barlow, M.J., & Sugermann, B.E.K., 2007, MNRAS, 375, 753

Fonfrıa, J.P.. Cernicharo, J., Richter, M.J., et al., 2008, ApJ, 673, 445

Frenklach, M., Clary, D.W., Gardiner, W.C. Jr, & Stein, S.E., 1984, 20th Symp. (Int.)on Combustion, The Combustion Institute, 887

Frenklach, M., & Warnatz, J., 1987, Combustion. Sci. Tech., 51, 265

Frenklach, M., & Feigelson, E.D., 1989, ApJ, 341, 372

Haynes, B.S., & Wagner, B.J., 1981, Prog. Energy Combust. Sci., 7, 229

Hony, S., Waters, L.B.F.M., & Tielens, A.G.G.M., 2002, A&A, 390, 533

Ivezic, J., & Elitzur, M., 1996, MNRAS, 279, 1019

Jager, C., Huisken, F., Mutschke, H., et al., 2009, ApJ, 696, 706

Jager, C., Mutschke, H., Henning, T., & Huisken, F., 2011, this volume

Justtanont, K., Decin, L., & Schoier, F.L., 2010 [arXiv:1007.1536]

Keady, J.J., Hall, D.N.B., & Ridgway, S.T., 1988, ApJ, 326, 832

Keady, J.J., & Ridgway, S.T., 1993, ApJ, 199, 214

Kifonidis, K., Plewa, T., Scheck, L., Janka, H.-Th., & Muller, E., 2006, A&A, 453, 661

Krestinin, A.V., Kislov, M.B., Raevskii, O.I., et al., 2000, Kinetics and Catalysis, 41, 90

Lambert, D., Rao. N.K., Pandey, G., & Evans, I.I., 2001, ApJ, 555, 925


Melnick, G., Neufeld, D.A., Ford, K.E., et al., 2001, Nature, 412, 160

Messenger, S., Amari, S. Gao, X., et al., 1998, ApJ, 502, 284

Miller, J.A., & Melius, C.F., 1992, Combustion and Flame, 91, 21

Miller, J.A., & Klippenstein, S.J., 2003, J. Phys. Chem. A, 107, 7783

Minutolo, F., Gambi, G., & D’Alessio, A., 1998, 27th Symp. (Int.) on Combustion, TheCombustion Institute, 1641

Mukherjee, J., Sarofim, A.F., & Longwell, J.P., 1994, Combustion and Flame, 96, 191

Neufeld, D.A., Gonzalez-Alfonso, E., Melnick, G., et al., 2010 [arXiv:1007.1235]

Nozawa, T., Kozasa, T., Taminaga, N., et al., 2010, ApJ, 713, 356

Olofsson, H., 2006, Rev. Modern Astron., 19, 75

Rao, N.K., & Lambert, D., 1993a, MNRAS, 263, L27

Rao, N.K., & Lambert, D., 1993b, AJ, 105, 1915

Rao, N.K., & Lambert, D., 2000, MNRAS, 313, L33

Richter, H. & Howard, J.B., 2000, Prog. Energy. Comb. Sci., 26, 565


Richter, H., & Howard, J.B., 2002, Phys. Chem. Chem. Phys., 4, 2038

Scherer, S., Just, Th., & Frank, P., 2000, Proc. Comb. Institute, 28, 1511

Smolders, K., Acke, B., Verhoelst, T., et al., 2010, A&A 514, L1

Speck, A., & Barlow, M.J., 1997, Ap&SS, 251, 115

Tang, W., Tranter, R.S., & Brezinsky, K., 2006, J. Phys. Chem. A, 110, 2165


Westmoreland, P.R., Dean, A.M., Howard, J.B., & Longwell, J.P., 1989, J. Phys. Chem.,93, 8171

Willacy, K., & Cherchneff, I., 1998, A&A, 330, 676

Willacy, K., 2004, ApJ, 600, L87

Zinner, E., 1997, AIPC, 402, 3


INSIGHTS INTO THE CONDENSATION OF PAHSIN THE ENVELOPE OF IRC +10216

L. Biennier1, H. Sabbah1, 2, S.J. Klippenstein3, V. Chandrasekaran1, 4,I.R. Sims1 and B.R. Rowe1

Abstract. The mechanisms of nucleation and growth of carbon dustparticles in circumstellar envelopes of carbon-rich stars in the red giantand AGB phases of their evolution are poorly understood. It has beenproposed that the transition of gas phase species to solid particles, isachieved by the formation of a critical nucleus composed of two PAHsheld together by van der Waals forces. Some insights into the validity ofthe nucleation of PAH molecules in the envelope can be gained throughthe investigation of the thermodynamics of dimers, representing thefirst stage towards condensation. We have performed experiments toidentify the temperature range over which small PAH clusters form insaturated uniform supersonic flows. The kinetics of the formation hasalso been investigated. The experimental data have been combinedwith theoretical calculations. We unambiguously demonstrate that theassociation of small PAHs such as pyrene (C16H10) is slower than thedestruction of the dimer in warm and hot environments such as IRC+10216. Our findings challenge a formation model based on the phys-ical stacking of small PAH units in circumstellar shells of carbon richstars.

1 Introduction

The outflows of carbon rich stars in their asymptotic giant branch phase (AGB)are the primary sources of carbon dust (Henning & Salama 1998). AGB stars

1 Institut de Physique de Rennes, Equipe: “Astrochimie Experimentale”, UMR CNRS 6251,Universite de Rennes 1, Campus de Beaulieu, 35042 Rennes Cedex, France2 Current address: Zarelab, Chemistry Department, Stanford University, CA 94305, USA3 Chemical Sciences and Engineering Division, Argonne National Laboratory, Argonne,IL 60439, USA4 Current address: Institut des Sciences Moleculaires, UMR CNRS 5255, UniversiteBordeaux 1, 33405 Talence, France



lose up to 80% of their original mass in the form of an envelope, with typical lossrates of 10−6 to 10−4 M� yr−1. There is limited observational evidence for theformation of dust in other sources, such as Supernovae (SNe) and novae, but theirrelative importance as sources of interstellar dust, are still uncertain (Jones 2005).

The determination of the physical conditions under which dust is generatedis critical. Estimates of the temperature and pressure of circumstellar envelopesstrongly depend on various model assumptions, such as mass loss rates, gas densityoutflow, and velocity. Observations have revealed that the envelope of evolved starsis huge, extending over 104 to 105 stellar radii. Close to the photosphere (<5 R�),the gas is dense (n ≥ 1010 molecule·cm−3) and warm (T ∼ 1500 K). As the matterflows away from the star, it expands (n ≤ 105 molecule·cm−3) and cools rapidlydown to 25 K at several hundreds of R� (Ziurys 2006).

The most refractory material composing the atmosphere, such as carbides, nu-cleates first in the innermost region, even inside the photosphere. Once formed,they act as nuclei sites for heterogeneous condensation of more volatile carbona-ceous molecules. C condensation is delayed until homogeneous nucleation can oc-cur at higher supersaturation levels farther out from the star (Bernatowicz et al.2006). Some complexity is added to this already debated picture, by the propa-gation of shock waves, produced by stellar pulsations, which greatly enhance thedensity and temperature just after the shock front. Because of the variety of phys-ical conditions, the chemistry varies substantially throughout the envelope. Inthe inner shell (< a few R�), the temperature and density are sufficiently highthat thermodynamic equilibrium governs molecular abundances of a number ofmolecules (Ziurys 2006). This assumption is confirmed by astronomical observa-tions in the infrared, which have shown that stable, closed-shell species such asCS, NH3, HCN, and HCCH have high abundances in the inner envelope of thewell studied IRC+10216 carbon star (≈10−3 to 10−6, relative to H2).

As the density falls with radius, interstellar UV photons penetrate the atmo-sphere and generate reactive species through photodissociation that subsequentlyparticipate in a complex gas-phase chemistry to produce daughter molecules inthe outer envelope. Interferometer maps of IRC +10216 reveal radicals such asCN, CCH, C3H, and C4H organized in shell-like structures near the envelope edge(Guelin et al. 1993).

It is generally accepted that the production pathway determines the result-ing structure, chemical composition, and morphology of the grains (Jager et al.2009). However, the formation pathways of carbonaceous grains from molecularcomponents and clusters, evoked by Jager in this volume (2010), are not wellunderstood. From a chemical viewpoint, the high abundance of hot acetylene(Fonfria et al. 2008) and the scarcity of oxydizing molecules in circumstellar en-velopes of carbon rich stars are expected to lead to the production of carbonaceousparticles through processes similar to the ones encountered in terrestrial low pres-sure flames. The analogy strongly suggests the presence of Polycyclic AromaticHydrocarbons (PAHs) as building blocks, intermediates or side products in thegeneration of carbon dust particles. Remarkably, the aromatic IR emission fea-tures, characteristic of PAHs, are generally not seen in the spectra of carbon-rich

L. Biennier et al.: Circumstellar PAHs 193

AGB stars, but are systematically detected in carbon-rich post-AGB objects andPlanetary Nebulae. Two interpretations are contending to explain the infraredemission features in the spectra of the latter sources. It can be the consequenceof efficient pumping arising from the increased effective temperature of the centralobject as it loses its remnant envelope and shifts to the blue in the Hertzsprung-Russell diagram on its way to becoming a white dwarf. Alternatively, it may bethe result of photochemical processing of hydrogenated amorphous carbon (HAC)material formed in the outflows. The ground-based and ISO/SWS detection of thearomatic infrared emission features in the spectra of the carbon-rich AGB star TUTau (Boersma et al. 2006), likely irradiated by its blue binary companion, doesnot provide an unambiguous argument in favor of either of the scenarios. Indepen-dent of the formation pathways actually followed, the presence of PAH featuresin the IR spectra of carbon-rich post-AGB objects and planetary nebulae (PNe)lends indirect support for the importance of late-type stars as sources for PAHs inthe interstellar medium (Tielens 2008).

Theoretical studies on the chemical pathways towards PAHs in circumstellarenvelopes are reviewed in this conference by Cherchneff (2010). Briefly, the for-mation of PAHs from small gaseous hydrocarbons in the circumstellar envelopeswas initially proposed by Frenklach & Feigelson (1989b). Their study was thefirst to introduce chemical kinetics in those atmospheres. The chemical schemeconsidered was largely inspired from combustion studies and relies heavily on thepresence of acetylene C2H2 and radicals such as C2H. Reactions involving oxygenwere removed, whereas benzyne C6H4 was introduced in the network. The domi-nant PAH formation pathway was identified as the hydrogen abstraction acetyleneaddition (HACA) route. Cherchneff revised and extended the scheme, explicitlyincluding radicals, such as C3H3 and aromatics up to C18H12 (Cherchneff et al.1992). In the study, Cherchneff et al. considered that the gas density was solow that collisions did not couple efficiently the vibrational and translational de-grees of freedom. As a consequence, PAH molecules did end up with vibrationaltemperatures significantly lower than gas kinetic temperature. Nevertheless, theconclusion was that PAH formation yields were very sensitive to gas density andtemperature, and that they were much smaller than values inferred from the ob-served dust content of late type carbon rich stellar envelopes. They suggested thatthe PAHs may form instead very close to the photosphere. Allain et al. furtherexplored the processes affecting PAH formation and growth (Allain et al. 1997).They carried out calculations of the PAH temperature by balancing heating viacollisions with gas particles and absorption of stellar emission and cooling via emis-sion from the rotational and vibrational levels. In their work they treated heatingprocesses simultaneously and found that low PAH temperatures enable an efficientformation and growth of PAHs in the circumstellar envelopes of C-stars. The calcu-lated PAH formation rate was large enough to support the idea that PAHs couldconstitute the first step in the formation of dust grains. The processes, eitherchemical or physical, that could explain the transition from these large moleculesto small dust grains, were however not investigated. Nucleation triggered by as-sociation of PAH molecules formed in the gas phase, considered as a key step in


controversial models of soot formation in flames (Frenklach 2002; Frenklach &Wang 1991), was initially discounted as the main mechanism of carbon dust pro-duction (Cadwell et al. 1994; Cherchneff et al. 1992; Frenklach et al. 1989a).Astrophysical models were favoring high temperature nucleation of mineral parti-cles in or near the photosphere, onto which PAH molecules could later condense(Cherchneff & Cau 1998). In an attempt to assess alternative pathways of dustformation, which could be potentially competing, Cherchneff (2000) investigatedthe formation of van der Waals bound PAH dimers in the envelopes, consideringspecies as small as benzene (C6H6) and up to coronene (C24H12). The conclusionwas that PAH dimers could act as dust nuclei in the inner shells (Cherchneff 2000).Cau (2002) later on included periodic shocks in the model and inferred that theamount of PAHs and dimers produced, considering aromatics containing up to7 rings, is not enough to explain the formation of carbon dust in the atmosphereof the IRC +10216 star, but that it can account for the formation of the smalldisordered cores of the kind observed in presolar grains (Cau 2002).

In this paper, we re-examine the condensation of PAH molecules in the en-velopes of carbon rich stars based on high level theoretical calculations supportedby state-of-the-art laboratory experiments related to the thermodynamics and ki-netics of the process. Our findings are applied to the envelope of the evolved redgiant, IRC +10216.

2 Combined experimental/theoretical approach

Laboratory experiments have been conducted to identify the temperature rangeunder which pyrene (C16H10) molecules dimerize (Sabbah et al. 2010). Our ex-perimental approach is briefly summarized below to provide the reader with someexperimental background (Sabbah et al. 2010). The experiments were performedusing a continuous flow CRESU apparatus (Canosa et al. 2008; Dupeyrat et al.1985), adapted for condensable species such as PAHs (Goulay et al. 2006). Thischemical reactor is designed to generate dense uniform flows using a Laval nozzle,over a wide range of temperatures (60–470 K) and containing high (supersatu-rated) concentrations of PAH vapors. The time evolution of the chemical speciespresent in the reactor is monitored by a time of flight mass spectrometer (TOFMS)equipped with a VUV laser for photoionization of the neutral reagents and prod-ucts. The onset of nucleation is explored for a set of flow temperatures. It isevidenced by the collapse of the PAH ion signal above a certain initial PAH con-centration. At these high degrees of supersaturation, dimers may be considered ascritical nuclei within the framework of classical nucleation theory. Once they areformed, they go on to react rapidly with more monomer to form trimers and largerclusters. After the identification of the temperature range over which nucleation isoccurring, we proceed to kinetics measurements on dimerization within this range,but at lower PAH monomer concentrations. This yields the apparent second orderrate coefficient for removal of PAH monomer, according to simple second-orderkinetics.

The experimental results are coupled with theoretical calculations that employcareful consideration of the intermolecular interaction energies and intermolecular


dynamics to estimate the binding energy, equilibrium constant, and rate coefficient(Sabbah et al. 2010). The theoretical analysis employs a sum of atom-atom poten-tials to describe the intermolecular interactions. The parameters in this potentialare designed to reproduce the equilibrium binding energies from high level calcula-tions for a range of small PAH dimers, while at the same time accurately modelingthe interactions for the large separations that span the transition state region forthe association. The analysis focuses on the rigid body dynamics of the two PAHmolecules, and employs transition state theory and trajectory simulations to ex-amine the high pressure addition kinetics. Predictions for the equilibrium constantare obtained from Monte Carlo integration of the classical phase integral repre-sentation of the intermolecular partition function. The intramolecular partitionfunctions of the monomer are assumed to be invariant to the dimerization process.These quantitative coupled anharmonic predictions for the equilibrium constantare used in the following analysis.

3 Astrophysical implications

The stable cluster structures of PAHs ranging from pyrene (C16H10) to circum-coronene (C54H18) have been explored theoretically recently (Rapacioli et al.2006). Stacking PAH molecules was found to yield the most stable motif. Theatomistic model developed in that work was then used in molecular dynamics sim-ulations to study the nucleation and evaporation of coronene (C24H12) clusters.Rapacioli et al. (2006) found that under cold interstellar conditions, most of thecollisions lead to cluster growth. They claimed that PAH clusters could repre-sent the missing link between small carbonaceous grains and free isolated PAHmolecules. Here we extend the exploration to the much warmer environment ofthe envelope of a late carbon star in which PAHs are supposedly generated andthe existence of PAH clusters postulated. We first describe below the physicalconditions characterizing the extensively studied, albeit not fully understood, en-velope of IRC+10216. This late carbon rich star is one of the brightest sources inthe infrared sky as a result of thermal re-emission by dust.

The modeling of infrared astronomical observations, led Fonfria et al. to dividethe envelope in concentric regions (Fonfria et al. 2008). They first distinguish theinner shell comprised between the stellar photosphere and about 5 R�. The outeredge corresponds to the first grain formation zone. Radiation pressure on grains,together with momentum coupling of gas and grains, strongly accelerate the gasup to vexp(r) ∼ 11 kms−1. In the intermediate envelope, the gas expands adiabat-ically until it reaches ∼20 R� where it is strongly accelerated for a second time upto its terminal velocity ∼14.5 km s−1. The second acceleration zone correspondsto the second dust formation layer.

In this work, we adopt the radial kinetic temperature profile derived fromthe best fit of observed C2H2 lines (Fonfria et al. 2008) with an effective stellartemperature, Teff ∼ 2330 K. The temperature of molecules such as PAHs in theextended atmosphere is governed by balancing heating and cooling processes. PAHmolecules are heated by (1) collisions with abundant H2 gas; and (2) absorption of


Fig. 1. Minimum pyrene abundance for condensing 10% of the monomers versus the dis-

tance from the central star IRC +10216 under a high molecular hydrogen density scenario

from Agundez & Cernicharo (2006). Molecular abundances of H2 (Agundez & Cernicharo

2006) and C2H2 (Fonfria et al. 2008) are indicated for comparison. The condensation is

governed by the PAH temperature which slowly decouples from the gas because of the

high number of collisions. The difference between the gas kinetic temperature and the

internal pyrene temperature initially null at the photosphere rises up to a maximum of

250 K at around 5 R� and then slowly decreases down to 200 K at 25 R�.

central star radiation, which peaks in the near infrared. As small PAH molecules donot absorb efficiently in this energy range, heating is therefore mainly ensured bycollisions with molecular hydrogen. Farther from the central star, the total densitydrops, collisions become less frequent and the vibrational temperature of the PAHsmay then decouple from the kinetic temperature of the gas. The main coolingmechanism has been identified as cascade emission of infrared photons. We followthe approach developed by Cherchneff et al. to derive the vibrational temperatureof pyrene. Adopting a spherical geometry and neglecting bipolar outflow cavities,the density profile, n(r), is given by the law of conservation of mass and thereforedepends linearly on the mass loss rate and the expansion velocity:

M = n(r)μmH vexp(r) 4πr2 (3.1)

with mH , the mass of the hydrogen atom, and μ ∼ 2 assuming that all hydrogenis molecular. Following the work of Agundez (Agundez et al. 2006), we considertwo different density profiles of the gas across the envelope hence bracketing therange of mass loss rates.

The equilibrium constant for pyrene dimerization (Sabbah et al. 2010) deter-mined from the combined experimental/theoretical work is used to calculate the


Fig. 2. Minimum pyrene abundance for condensing 10% of the monomers versus the

distance from the central star IRC +10216 under a low molecular hydrogen density

scenario from Agundez & Cernicharo (2006). Molecular abundance of H2 (Agundez &

Cernicharo 2006) is indicated. The condensation is governed by the PAH temperature

which decouples from the gas kinetic temperature very close to the star. At the distance

of 5 R�, the pyrene internal temperature has already dropped below 100 K.

minimum PAH number density for condensing 10% of the PAH monomers at agiven temperature. The resulting minimum abundances are plotted in Figures 1and 2 along with the molecular hydrogen and acetylene density profiles.

The most striking feature is the high sensitivity of the minimum pyrene abun-dance to the total gas density. As soon as the gas density drops, collisions arenot enough to keep the PAH molecules hot. The PAH vibrational temperaturestarts to decouple from the gas kinetics temperature, and therefore the associa-tion between two cooler PAH molecules becomes faster than the destruction ofthe dimer. When considering a high relative abundance of 10−9 for each of thePAHs, the plots show that condensation may occur >5 R� and 35 R� under thelow and high density scenarios, respectively. Based on these observations, we firstconclude that under the high gas density scenario, condensation of small PAHs oc-curs too far from the star (>35 R�) to represent a key mechanism in the formationof carbon dust particles.

The low gas density scenario seems more favorable. We look now at the kineticsof the association process to extract some time constraints. Collisions betweencold PAH molecules above the edge of the inner shell lead to an energized complexwhich either relaxes by emission of infrared photons, relaxes by collision with a


third body, or breaks apart. Calculations have revealed a surprising decouplingbetween the inter- and intra-molecular modes in the PAH complex (Sabbah et al.2010). This means that under low pressure environments the excess collisionalenergy may not be that easily evacuated through infrared emission and may rathercause the complex to break apart, hence reducing its lifetime. Nevertheless, thehypothesis of a 100% sticking efficiency provides an upper limit to the lifetime ofthe PAH complex. A simple calculation then gives a lower limit for the growthtime – corresponding to the inverse of the product of the capture rate and thePAH abundance – of ∼109 s. This characteristic growth time is 20 times longerthan the period of the shock, ∼630 days, that strongly affects the inner shell ofthe carbon star. Based on this kinetics argument, one can then conclude thatthe formation of PAH clusters is slow and will then be perturbed by the rise intemperature of the environment caused by the propagation of the shock.

4 Conclusion

We have re-examined the condensation of small PAHs in the envelope of IRC+10216.As the distance from the star grows, the gas density drops and the temperatureof the PAH molecules decouples from the gas kinetic temperature. Cold PAHsmolecules are then more likely to stick to each other and form a dimer. The ther-modynamics of pyrene dimers in this environment is explored using the equilibriumconstant of the dimer derived from the combined experimental/theoretical work.

Our results show that a low density of gas is necessary for small PAHs such aspyrene to condense at the edge of the inner shell of the star. However, even then thegrowth of these clusters is expected to be slow and therefore seriously hinderedby periodic shocks induced by stellar pulsations. This rules out clustering ofsmall PAHs as the main mechanism for generation of carbon particles in theseenvironments. Our conclusions do not contradict the findings of Bernatowitcz(2006) as heterogeneous condensation of carbonaceous molecules on carbides isstill possible. Alternatively, larger PAHs could stack efficiently under conditionsprevailing in high temperatures environments. The chemical pathway leading tothese large aromatic units remains to be examined.

In the laboratory, there are a number of challenges to take up to explore thekinetics of dimerization of larger PAHs than pyrene. The most difficult one isprobably to evaporate these condensable species and mix them with the buffer gasto produce a uniform continuous flow with a sufficient (and known) amount ofreactant, stable over time. An alternative such as laser desorption is not suitablebecause, although the technique can gently vaporize molecules, the amount ofmolecules desorbed is unknown and it generally relies on pulsed laser systems.

Future studies will then rather extend the theoretical approach developed herefor pyrene to larger species more prone to condensation.

A more complete study would include periodic shocks which boost the density(and therefore the number of collisions) and cause a rise in temperature of the


medium. Other routes based on chemical growth should be explored by chemicalkinetics models.

The authors acknowledge funding from the Indo-French Centre for the Promotion of AdvancedResearch (Project number 3405-3). This work is also supported by the French Programmesof Physique Stellaire and of Physique et Chimie du Milieu Interstellaire. The theoretical work(S.J.K.) has been supported by NASA’s Planetary Atmospheres Program through grantNNH09AK24I and by the US Department of Energy, Office of Basic Energy Sciences, Divisionof Chemical Sciences, Geosciences, and Biosciences under Contract No. DE-AC02-06CH11357.

References

Agundez, M., & Cernicharo, J., 2006, ApJ, 650, 374

Allain, T., Sedlmayr, E., & Leach, S., 1997, A&A, 323, 163

Bernatowicz, T., Croat, T.K., & Daulton, T.L., 2006, MESS, 943, 109

Boersma, C., Hony, S., & Tielens, A., 2006, A&A, 447, 213

Cadwell, B.J., Wang, H., Feigelson, E.D., & Frenklach, M., 1994, ApJ, 429, 285

Canosa, A., Goulay, F., Sims, I.R., & Rowe, B.R., 2008, In Low Temperatures and ColdMolecules, ed. IWM Smith, 55 (Singapore: World Scientific)

Cau, P., 2002, A&A, 392, 203

Cherchneff, I., Barker, J.R., & Tielens, A., 1992, ApJ, 401, 269

Cherchneff, I., & Cau, P., 1998, Proceedings of the Asymptotic Giant Branch Stars 191thIAU Symp., 251

Cherchneff, I., 2000, The Carbon Star Phenomenon, Proceedings of the 177th IAU Symp.

Cherchneff, I., 2011, Proceedings of PAHs and the Universe Conference, this volume

Dupeyrat, G., Marquette, J.B., & Rowe, B.R., 1985, In Physics of Fluids, 1273

Fonfria, J.P., Cernicharo, J., Richter, M.J., & Lacy, J.H., 2008, ApJ, 673, 445

Frenklach, M., 2002, Physical Chemistry Chemical Physics, 4, 2028

Frenklach, M., Carmer, C.S., & Feigelson, E.D., 1989, Nature, 339, 196


Frenklach, M., & Wang, H., 1991, Proceedings of the Combustion Institute, 23, 1559

Goulay, F., Rebrion-Rowe, C., Biennier, L., et al., 2006, JPC A, 110, 3132

Guelin, M., Lucas, R., & Cernicharo, J., 1993, A&A, 280, L19

Henning, T., & Salama, F., 1998, Science, 282, 2204

Jager, C., Huisken, F., Mutschke, H., Jansa, I.L., & Henning, T.H., 2009, ApJ, 696, 706

Jager, C., Mutschke, H., Henning, T., & Huisken, F., 2011, Proceedings of PAHs andthe Universe Conference, this volume

Jones, A.P., 2005, In Proceedings of the Dusty and Molecular Universe: a Prelude toHerschel and Alma, ed. A Wilson, 239 (Ag Noordwijk: Esa Publications DivisionC/O Estec), 2200

McCabe, E.M., 1982, MNRAS, 200, 71

Rapacioli, M., Calvo, F., Spiegelman, F., Joblin, C., & Wales, D.J., 2005, JPC A, 109,2487

Sabbah, H., Biennier, L., Klippenstein, S.J., Sims, I.R., & Rowe, B.R., 2010, JPC Lett.,2962

Tielens, A., 2008, ARA&A, 46, 289

Ziurys, L.M., 2006, PNAS, 103, 12274


FORMATION AND EVOLUTION OF CIRCUMSTELLARAND INTERSTELLAR PAHS: A LABORATORY STUDY

C.S. Contreras1, C.L. Ricketts1 and F. Salama1

Abstract. Studies of dust analogs formed from hydrocarbon (CH4,C2H2, C2H4, C2H6) and PAH precursors have been performed us-ing a new facility that we have developed to simulate interstellar andcircumstellar processes. The species formed in a plasma are detected,characterized and monitored in situ with high-sensitivity techniques,which provide both spectroscopic and ion mass information. Fromthese measurements we derive information on the nature, the size andthe structure of dust particles, as well as a better understanding of thegrowth and destruction processes of extraterrestrial dust.

1 Introduction

The study of the formation, growth and destruction mechanisms of cosmic dustgrains is essential for a correct understanding of the physical and chemical evo-lution of the interstellar medium (ISM). Dust particles span a continuous sizedistribution from large molecules to nanometer-sized particles to μm-sized parti-cles and influence many processes in the evolution of the ISM such as the energybalance through the photoelectric effect, the ionization balance through recombi-nation with electrons and ions, and the chemical composition of molecular clouds(Tielens 2005, see contributions by Verstraete and Bierbaum et al. elsewhere inthis volume). The carbonaceous component of cosmic dust plays an importantrole in the universe (Henning & Salama 1998). Carbon particles are thought to beprimarily formed in the outflow of carbon stars, through a combustion-like processwhere small carbon chains (e.g., acetylene) form polycyclic aromatic hydrocarbons(PAHs) that nucleate into larger-size PAHs and, ultimately, into nanoparticles(Frenklach et al. 1989; Pascoli & Polleux 2000; Jager et al. elsewhere in this vol-ume). According to the model, that was developed to account for the formationof carbonaceous meteoritic materials, nucleation occurs above 2000 K, followed bythe growth of amorphous carbon on the condensation nuclei in the 2000–1500 K

1 NASA Ames Research Center, Moffett Field, CA, USA



temperature range. As the temperature falls to around 1100 K, aromatic moleculesbegin to form in the gas phase and condense onto the growing particles forminggraphitic microstructures that will ultimately aggregate into larger structures suchas seen in soot formation (Cherchneff et al. elsewhere in this volume).

Although a large amount of laboratory data exists on the formation and theproperties of solid dust analogs and in particular of Si-based materials (e.g.,Michael et al. 2003), very little is known, about the specific properties (structure,size and spectral signatures) and the formation mechanisms of the key intermedi-ate range of the size distribution (nanoparticles) of cosmic dust. The scarcity ofrelevant laboratory data on the transition from gas-phase molecules to solid grainsis due to the difficulty of forming and isolating these large species and in trackingtheir evolutionary path under astrophysically realistic laboratory conditions.

Here we describe new laboratory studies that have been performed to addressthis key issue. First we describe the experimental set-up that has been developedto measure the formation and destruction mechanisms of carbonaceous grains frommolecular precursors in an environment with key physical parameters expected toexist in circumstellar regions where molecular formation is observed. We thenpresent and discuss the preliminary results that were obtained in the laboratoryand conclude with a discussion of the implications of these new results.

2 Experimental

2.1 Experimental apparatus

The Cosmic Dust Simulation Chamber (CDSC) was recently set up in our labo-ratory for the study of cosmic analogs (Fig. 1). The CDSC has been previouslyused to obtain the first electronic spectra of PAHs and small nanoparticles underconditions that realistically simulate astrophysical environments (for a review seeSalama 2008). It combines three complementary techniques: Cavity Ring DownSpectroscopy (CRDS), Pulsed Discharge Nozzle (PDN) Expansion and ReflectronTime-of-Flight Mass Spectrometry (ReTOF-MS) providing both spectroscopic andion mass information for the identification of the species (large molecules, nano-sized grains) that are formed in the plasma source.

The PDN allows the study of molecules, ions and particles isolated in the gasphase. The species are generated in the hot, confined, plasma from carbonaceousprecursors and are suddenly frozen by the expansion (Jost 1996) providing a veryefficient cooling over short distances (a few mm) and an ideal simulation tool forastrophysical environments. The current geometry of the PDN leads to a residencetime of a few microseconds for the particles in the active region of the discharge.The plasma expansion generated by the PDN is well characterized (Remy et al.2003; Broks et al. 2005) and is probed by CRDS several mm downstream witha sub-ppm to ppm sensitivity. CRDS is based on a Nd:YAG pumped tunablepulsed dye laser for the injection of UV-visible photons into a high-finesse cavity(Biennier et al. 2003; Tan & Salama 2005). For the ReTOF-MS, in order to

C.S. Contreras et al.: Formation and Evolution of PAHs 203

Fig. 1. The Cosmic Dust Simulation Chamber (CDSC) consists of a Pulsed Discharge

Nozzle coupled to a Cavity Ring Down Spectrometer apparatus and to an orthogonal,

Reflectron time-of-flight mass spectrometer (PDN-CRDS-oReTOFMS).

collect and analyze the ions formed in PDN plasma, a portion of the expansion isskimmed through a 2 mm orifice which is attached to a long tube that leads to asecond aperture of 2 mm. The second aperture is polarized (–30 V) in order toattract the positive ions from the plasma. A repelling pulse of +200 V is appliedwhen the positively-charged ions have arrived in the extraction region; the ions aresteered orthogonally to their original flight path, to minimize noise, and guidedinto the free flight tube of the ReTOF-MS. The ion pulse is then back reflectedwith the Reflectron optics toward the MCP detector. The ions are dispersed intime according to their velocity (proportional to their mass-to-charge ratio, m/z)so the discrete packets of different m/z ions are detected (Ricketts et al. 2011).

2.2 Experimental procedures

The laboratory species that are generated in the expanding discharge plasma ex-perience a strong temperature gradient that ranges from a few thousand degreesKelvin to 100 K over a short distance (1.5 – 2.5 mm) from the edge of the plasmachamber to the probing zone where the particles are detected (Remy et al. 2003;Broks et al. 2005; Biennier et al. 2006). This temperature domain corresponds tothe domain where aromatic molecules begin to form in the gas phase and condenseonto the growing particles forming graphitic microstructures that will ultimatelyaggregate into larger soot-like structures as we observe. This temperature varia-tion is similar to the temperature variation observed from the surface of a carbonstar to the edges of the cooling stellar outflow envelope (Pascoli & Polleux 2000).

Plasma discharges were conducted with small hydrocarbons, methane, ethane,ethylene, and acetylene representative of the alkane, alkene, and alkyne groups, re-spectively. The formation of larger molecules and carbon grains was probed by theaddition of small PAH precursors (naphthalene, 1-methylnaphthalene,


Fig. 2. Representative mass spectra for plasma discharges into Argon seeded with hydro-

carbons and/or PAHs. Left: hydrocarbons in Argon; Right: mixtures of Hydrocarbons

and Acenaphthene (154 m/z) in Argon (close-up of the region for the Acenaphthene ion).

acenaphthene) into the hydrocarbon-seeded Ar gas mixture. In all experiments,gases are seeded into Ar and expanded through the PDN. Solid and liquid PAHsamples are held in the bottom reservoir of the PDN and are heated to increasetheir vapor pressure. In typical experiments, the high voltage discharge is 300 μslong and is applied within a gas pulse that lasts 1.2 ms. The ReTOF-MS extrac-tion pulse event is timed to coincide with the discharge and the time it takes theions to reach the mass spectrometer. The extraction pulse occurred 200 μs afterthe discharge.

3 Preliminary results

The spectra of the discharge products are shown in Figure 2.

3.1 Linear hydrocarbon precursors:

3.1.1 Alkanes

In the case of methane, a 1-C alkane, carbon build up leads to structures with2, 3 and 4 carbons. The highest mass ion detected is C4H

+6 . Dehydrogenation

(from CH3 to bare C) is also observed. Carbon build up is also observed in thecase of ethane, a 2-C atom alkane, with structures with 2, 3 and 4 carbons. Thehighest mass ion detected is C4H

+7 . Dehydrogenation (from C2H5 to C2) as well


as fragmentation into 1-C structures is observed. These results seem to indicatethat the 1-C alkane precursor (CH4) exhibits a higher reactivity.

3.1.2 Comparison of the alkane, alkene, and alkyne hydrocarbon classes

Ethylene and acetylene, both unsaturated hydrocarbons, show greater reactivityand propensity to form larger mass ions than ethane. In the case of ethylene,structures up to 5-C atoms are detected and the highest mass ion being C5H

+9 .

In the case of acetylene, structures up to 6-C atoms are found with the highestmass ion detected is C6H

+5 . It should be noted that acetylene is routinely stored

in acetone and that some of the peaks in the mass spectrum (e.g. 58 and 59 m/z)could be due to oxygen containing hydrocarbons (acetone and protonated acetoneions, respectively). The lower reactivity of the alkane, results in lower mass ionsformed in the plasma discharge.

3.2 Polycyclic aromatic hydrocarbons

As a basis for the study of the formation of larger PAHs, the naphthalene moietywas used and the reactivity of hydrocarbon-substituted naphthalene was studied.Experiments with naphthalene (C10H8), 1-methylnaphthalene (C11H10) and ace-naphthene (C12H10) were performed. While no evidence for bond formation wasfound, and therefore no ions with a higher m/z than the PAH precursors were de-tected in these preliminary runs, an interesting fragmentation pattern was observedto be distinct for each PAH molecule. For example, in the case of naphthalene, theformation of the C8H

+6 ion corresponds to a loss of C2H2 from the precursor. The

detection of the C4H+4 and its dehydrogenated ions (C4H

+3 , C4H

+2 ) demonstrates

the loss of C6Hx fragments, C6H4, C6H5, and C6H6, respectively. The hydrogenloss observed for naphthalene seems to occur in a sequential manner, since boththe C10H

+7 and C10H

+6 ions are observed. In the case of 1-methylnaphthalene,

the fragmentation process seems to be distinct to that of naphthalene. There isa single H loss (C11H

+9 , 141 m/z), and instead of a sequential loss of hydrogen,

no peak is observed at 140 m/z. A corresponding fragment ion (C9H+7 ) is ob-

served at 115 m/z, which is a total loss of C2H3. In the acenaphthene spectrum,two clear, sequential hydrogen losses are observed. One possible explanation forthis observation is that the hydrogen loss most likely occur at the two adjacent,saturated carbons, thereby forming a double bond between them, with no otherfragment ions detected.

3.3 PAHs with hydrocarbons

Plasma experiments with both hydrocarbons and PAHs seeded in Argon indicateminimal reactivity between the two carbon containing species. A notable exceptionis the formation of the doubly-charged C2H

2+3 ion (13.5 m/z), which only appears

in experiments with mixed precursors. With a few exceptions, most of the peaksthat are detected in this case follow the trend found for the hydrocarbon-only ex-periments, with similar relative peaks intensities. PAH molecular ion abundances


(128, 142, 154 m/z for naphthalene, 1-methylnaphthalene, and acenaphthene, re-spectively) increase when mixed with the hydrocarbons, as opposed to PAHs inArgon experiments. Follow-up studies are on-going to identify the processes re-sponsible for the observed ionization yields.

4 Conclusions

These preliminary experiments have demonstrated that adding a ReTOF-MS tothe Cosmic Dust Simulation Chamber (CDSC) constitutes a powerful tool to probethe formation and fragmentation processes of laboratory analogs of interstellarand circumstellar dust. Experiments with hydrocarbon molecule precursors haveindicated the formation of larger species through chain growth in all the casesprobed so far. The formation of doubly charged ions in the plasma probablycontribute to enhancing the reactivity in the plasma chamber. All experimentsshow fragmentation of the precursors, with losses of H , CH , C2H2, among otherfragments specific to each hydrocarbon/PAH combination. The PAHs naphthaleneand acenaphthene show subsequent hydrogen loss, with naphthalene dissociatinginto smaller hydrocarbon fragments. The discharge has been known to inducebreakdown of naphthalene in the plasma when the applied voltage on the electrodesis higher than –500 V (Remy et al. 2003). For acenaphthene, the loss of twohydrogen atoms most likely occur at the C-C bond that is not part of the aromaticnetwork. Each carbon may lose a hydrogen, forming a double bond between thecarbons. The resulting end product is the acenaphthylene (C12H

+8 ) ion, that is

slightly larger than naphthalene. 1-Methylnaphthalene has a total loss of threehydrogens, where two of the hydrogens are lost in one step. The loss of twocarbons to form C9H

+7 , and no evidence for the naphthalene cation (127 m/z) in

the mass spectrum, tend to indicate that there is ring opening and fragmentationalong the ring. No other fragments are detected, possibly due to the stabilizationeffect of ring opening. There are indicators that multiply-charged ions are beingproduced in the plasma discharge as indicated by the detection of the C2H

2+3

(13.5 m/z) and C2H2+5 (14.5 m/z) ions. While C2H

2+5 is only detected when

ethane is present, C2H2+3 is detected in most cases. These observations seem to

point to the fact that the presence of the PAHs may facilitate the formation ofthe doubly-charged species between ions and neutrals from the hydrocarbon gas.More experiments are being performed to further explore these reactions.

This research was supported by the NASA APRA and Cosmochemistry Programs of the ScienceMission Directorate. C.S. Contreras and C.L. Ricketts hold NPP fellowships at NASA-AmesResearch Center. The authors wish to acknowledge fruitful discussions with L. Biennier and theoutstanding technical support of R. Walker.


References

Biennier, L., Benidar, A., & Salama, F., 2006, Chem. Phys., 326, 445

Biennier, L., Salama, F., Allamandola, L., & Scherer, J., 2003, J. Chem. Phys., 118, 7863

Broks, B., Brok, W., Remy, J., et al., 2005, Phys. Rev., E., 71, 36409

Frenklach, M., Carmer, C., & Feigelson, E., 1989, Nature, 339, 196


Jost, R., 1996, in Low Temperature Molecular Spectroscopy, ed. R. Fausto, 483 (KluwerAcademic Publishers, Dordrecht, NL), 249

Michael, B., Joseph III, A., & Lilleleht, L., 2003, ApJ, 590, 579

Pascoli, G., & Polleux, A., 2000, A&A, 359, 799

Remy, J., Biennier, L., & Salama, F., 2003, Plasma Sources Sci. Technol., 12, 295

Ricketts, C.L., Contreras, C.S., & Salama, F., 2011, Int. J. Mass Spectrom. Ion Processes,300, 26

Salama, F., 2008, in Organic Matter in Space, IAU Proceedings, ed. Kwok, & Sandford(Cambridge University Press), 357, and references therein

Tan, X., & Salama, F., 2005, J. Chem. Phys., 122, 084318

Tielens, A., 2005, The Physics and Chemistry of the Interstellar Medium (CambridgeUniv Press, Cambridge, UK)


CONFIRMATION OF C60 IN THE REFLECTION NEBULANGC 7023

K. Sellgren1, M.W. Werner2, J.G. Ingalls3, J.D.T. Smith4,T.M. Carleton5 and C. Joblin6, 7

Abstract. The fullerene C60 has four infrared active vibrational transi-tions at 7.0, 8.5, 17.5 and 18.9 μm. We have previously observed emis-sion features at 17.4 and 18.9 μm in the reflection nebula NGC 7023and demonstrated spatial correlations suggestive of a common origin.We now confirm the identification of these features with C60 by detect-ing a third emission feature at 7.04 ± 0.05 μm at a position of strong18.9 μm emission in NGC 7023. We also report the detection of thesethree features in the reflection nebula NGC 2023. We show with spec-troscopic mapping of NGC 7023 that the 18.9 μm feature peaks onthe central star, that the 16.4 μm emission feature due to PAHs peaksbetween the star and a nearby photodissociation front, and that the17.4 μm feature is a blend of a PAH feature and C60. The derivedC60 abundance is consistent with that from previous upper limits andpossible fullerene detections in the interstellar medium.

1 Introduction

Fullerenes are cage-like molecules (spheroidal or ellipsoidal) of pure carbon, suchas C60, C70, C76, and C84. C60 is the most stable fullerene and can account for upto 50% of the mass of fullerenes generated in the laboratory (Kroto et al. 1985).Foing & Ehrenfreund (1994) propose that two diffuse interstellar bands at 958 and963 nm are due to singly ionized C60, or C+

60, although the identification is debated

1 Department of Astronomy, Ohio State University2 Jet Propulsion Laboratory, California Institute of Technology3 Spitzer Science Center, California Institute of Technology4 Ritter Astrophysical Research Center, University of Toledo5 University of Arizona6 Universite de Toulouse, UPS, CESR7 CNRS



(Maier 1994; Jenniskens et al. 1997). Misawa et al. (2009) attribute additionaldiffuse interstellar bands at 902, 921, and 926 nm to C+

60. Observational evidencefor neutral fullerenes, however, has been elusive to date. No fullerenes have yetbeen found towards carbon-rich post-AGB stars (Somerville & Bellis 1989; Snow& Seab 1989), carbon stars (Clayton et al. 1995; Nuccitelli et al. 2005), or RCrB stars (Clayton et al. 1995; Lambert et al. 2001). Similarly, no neutralfullerenes have yet been found in the diffuse interstellar medium (Snow & Seab1989; Herbig 2000), dense molecular clouds (Nuccitelli et al. 2005), or at thephotodissociation front in the reflection nebula NGC 7023 (Moutou et al. 1999).

C60 has four infrared-active vibrational transitions, at 7.0, 8.5, 17.4, and 18.9 μm(Frum et al. 1991; Sogoshi et al. 2000). We tentatively identified the 17.4 and18.9 μm interstellar emission features in the reflection nebula NGC 7023 as due toC60 (Werner et al. 2004b; Sellgren et al. 2007).

We report here the detection of the C60 feature at 7.04 ± 0.05 μm in NGC 7023.We also report the detection of C60 features at 7.04, 17.4 and 18.9 μm in a sec-ond reflection nebula, NGC 2023. The C60 8.5 μm feature is too blended withthe strong 8.6 μm PAH feature to yield a useful limit. We have found earlier(Sellgren et al. 2007) that the 18.9 μm emission feature in NGC 7023 has a spatialdistribution distinct from that of the 16.4 μm emission feature attributed to poly-cyclic aromatic hydrocarbons (PAHs). We now compare the spatial distributionsof the 16.4, 17.4, and 18.9 μm emission features in NGC 7023, and find additionalsuppport for the C60 identification.

2 Results

We have used the Spitzer Space Telescope (Werner et al. 2004a) with theInfrared Spectrograph (IRS; Houck et al. 2004) to obtain spectra of NGC 7023 andNGC 2023. We used the short-wavelength low-resolution module SL (5–14 μm;Δλ/λ = 60–120) and the long-wavelength low-resolution module LL2 (14–20 μm;Δλ/λ = 60–120). Nebular positions were chosen for a strong ratio of the 18.9 μmfeature relative to the 16.4 μm PAH feature. We also made a spectroscopic mapwith the LL2 module of NGC 7023.

We show our 14–20 μm spectrum of NGC 2023 in Figure 1. We mark thewavelengths of C60 lines at 17.4 and 18.9 μm, of a PAH feature at 16.4 μm, andof H2 emission at 17.0 μm.

We show our 5–9 μm spectra of NGC 7023 in Figure 2. We clearly detectan emission feature at 7.04 ± 0.05 μm. We detect the 7.04 μm feature also inNGC 2023. This feature is coincident, within the uncertainties, with the wave-length of the expected C60 line. We highlight this emission feature by fitting the5–9 μm spectrum with a blend of PAH emission features in addition to the newemission feature at 7.04 μm. We perform this model fit using PAHFIT (Smithet al. 2007a).

In our previous long-slit spectroscopic investigation of NGC 7023 (Sellgrenet al. 2007), we found that the 18.9 μm feature peaks closer to the central starthan either neutral PAHs or ionized PAHs. We now use the LL2 spectroscopic

K. Sellgren et al.: C60 in Reflection Nebulae 211

14 15 16 17 18 19 20Wavelength (μm)

0.0

0.5

1.0

1.5

2.0

NG

C 2

023

Pos

ition

1 F

ν (Jy

)

C60+PAH

C60

H2

PAH

Fig. 1. Spitzer-IRS spectrum of NGC 2023 (solid black histogram), obtained with the

long-wavelength low-resolution module (LL2; 14–20 μm; λ/Δλ = 60–130). We mark the

wavelengths of C60 at 17.4 and 18.9 μm, a PAH feature at 16.4 μm, and H2 emission at

17.0 μm (vertical solid lines).

map extracted in NGC 7023 to illustrate this point more clearly in Figure 3.The 18.9 μm emission is clearly centered on the star. By contrast, the 16.4 μmPAH emission peaks outside the region of maximum 18.9 μm emission, in a layerbetween the star and the molecular cloud. The photodissociation front at theUV-illuminated front surface of the molecular cloud is delineated by fluorescentH2 emission at 17.0 μm.

We show an image of the 17.4 μm emission from NGC 7023 in Figure 3, overlaidwith contours of 18.9 μm and 16.4 μm emission. The 17.4 μm emission clearlyshows one peak on the central star, coincident with the 18.9 μm C60 emission, anda second peak co-spatial with the 16.4 μm PAH emission. We conclude that theobserved 17.4 μm emission feature is a blend of a PAH feature at 17.4 μm, whosespatial distribution follows that of the 16.4 μm PAH feature, and of 17.4 μm C60

emission.

3 Discussion

We have detected three lines due to C60, in two reflection nebulae. This is thefirst detection of neutral C60 in space.

Theorists have suggested that fullerenes might form around stars with carbon-rich atmospheres, such as carbon stars, cool carbon-rich Wolf-Rayet (WC) stars,and C-rich, H-poor R Cr B stars (Kroto & Jura 1992; Goeres & Sedlmayr 1992;Cherchneff et al. 2000; Pascoli & Polleux 2000). Fullerenes may also form as


Fig. 2. Spitzer-IRS 5–9 μm spectrum of NGC 7023, obtained with the short-wavelength

low-resolution module (SL; λ/Δλ = 60–120; solid black histogram). The wavelength

of C60 at 7.0 μm is marked (vertical solid magenta line). We show the individual con-

tributions of PAH features at 5.3, 5.7, 6.2, 6.4, 6.7, 7.4, 7.6, 7.8, 8.3, and 8.6 μm to

the spectrum, by decomposing the spectrum with PAHFIT (Smith et al. 2007a) and

then overplotting the Lorentzian profile of each feature (thin solid black curves). The

Lorentzian fit to the C60 feature we detect at 7.04 ± 0.05 μm is highlighted (thick solid

magenta curve).

part of the carbon-rich grain condensation process known to occur in the ejecta ofType II supernovae (Clayton et al. 2001). Fullerenes might form via interstellargas-phase chemistry in dense and diffuse molecular clouds (Bettens & Herbst 1996,1997). Hydrogenated amorphous carbon grains in the interstellar medium may alsodecompose after interstellar shocks into fullerenes and PAHs (Scott et al. 1997).

We compare the relative amounts of stellar flux absorbed and re-radiated byfullerenes and PAHs to derive the abundance of C60 in NGC 7023 and NGC 2023.First we compare the sum of the intensities of the 7.0, 17.4, and 18.9 μm features,assumed to be due to C60, to the sum of all other infrared emission features at5–20 μm, assumed to be due to PAHs. We ignore potential visual fluorescenceby C60 or PAHs, do not include any potential 8.5 μm C60 emission, and do notattempt to correct the 17.4 μm C60 emission for any PAH contribution. We usedPAHFIT (Smith et al. 2007a) to find that the ratio of C60 to PAH emission is0.01–0.03 in NGC 7023 and NGC 2023. We then compare the amount of UVstarlight per C atom absorbed by C60 (Yasumatsu et al. 1996; Yagi et al. 2009).and PAHs (Li & Draine 2001). We assume that 9–18% of interstellar carbonis in PAHs (Joblin et al. 1992; Tielens 2008). Our preliminary estimate of thepercentage of interstellar carbon contained in C60, p(C60), is 0.1–0.6%.

K. Sellgren et al.: C60 in Reflection Nebulae 213

Fig. 3. Spectral map of NGC 7023, obtained with the Spitzer IRS 15–20 μm LL2 long-

slit module, extracted and analyzed using CUBISM (Smith et al. 2007b). Each pixel

is 5.1′′ × 5.1′′. Left: three-color image of NGC 7023, with the 18.9 μm C60 feature in

red, the 16.4 μm PAH feature in green, and the 17.0 μm H2 line in blue. The position

of the central star is marked with a cross. The 18.9 μm C60 feature peaks on the central

star. The front surface of the UV-illuminated molecular cloud is traced by H2 emission,

while the 16.4 μm PAH emission peaks between the C60 and H2 emission. Right: gray-

scale image of NGC 7023 in the 17.4 μm feature, overlaid with contours of 18.9 μm C60

emission (red), and contours of 16.4 μm PAH feature emission (green). The observed

17.4 μm emission has two spatial components, one co-spatial with C60 emission and the

other co-spatial with PAH emission. We conclude that the 17.4 μm emission feature is

due to a blend of PAH and C60 emission.

Our preliminary estimate of p(C60) is consistent with published estimates ofthe percentage of carbon in C+

60 in diffuse clouds (0.1–0.9%; Foing & Ehrenfreund1994; Herbig 2000). It is also consistent with previous upper limits on p(C60)towards R Coronae Borealis, massive young stellar objects, and NGC 7023 (<0.3–0.6%; Moutou et al. 1999; Nuccitelli et al. 2005).

4 Conclusions

We confirm our previous suggestion that emission features at 17.4 and 18.9 μm inNGC 7023 are due to C60, by detecting a third C60 line at 7.04 μm. We detect the7.04, 17.4, and 18.9 μm C60 lines in NGC 2023 also. The spatial distribution of the17.4 μm feature in NGC 7023, compared to the spatial distributions of 16.4 μmPAH and 18.9 μm C60 emission, suggests that the observed 17.4 μm feature is asuperposition of two lines, one due to PAHs and one due to C60. The abundancewe derive for C60 is consistent with previous upper limits on C60 and previousmeasurements of lines identified with C+

60. Our observations are the first firmdetection of C60 in space.

This work is based on observations made with the Spitzer Space Telescope,which is operated by the Jet Propulsion Laboratory, California Institute of


Technology under a contract with NASA. Support for this work was providedby NASA through an award issued by JPL/Caltech.

References

Bettens, R.P.A., & Herbst, E., 1997, ApJ, 478, 585

Bettens, R.P.A., & Herbst, E., 1996, ApJ, 468, 686

Cherchneff, I., Le Teuff, Y.H., Williams, P.M., & Tielens, A.G.G.M., 2000, A&A, 357,572

Clayton, D.D., Deneault, E.A.-N., & Meyer, B.S., 2001, ApJ, 562, 480

Clayton, G.C., Kelly, D.M., Lacy, J.H., et al., 1995, AJ, 109, 2096

Foing, B.H., & Ehrenfreund, P., 1994, Nature, 369, 296

Frum, C.I., Engleman, R.J., Hedderich, H.G., et al., 1991, Chem. Phys. Lett., 176, 504

Goeres, A., & Sedlmayr, E., 1992, A&A, 265, 216

Herbig, G.H., 2000, ApJ, 542, 334

Houck, et al., 2004, ApJS, 154, 18

Jenniskens, P., Mulas, G., Porceddu, I., & Benvenuti, P., 1997, A&A, 327, 337

Kroto, H.W., Heath, J.R., O’Brien, S.C., Curl, R.F., & Smalley, R.E., 1985, Nature, 318,162


Kroto, H.W., & Jura, M., 1992, A&A, 263, 275

Lambert, D.L., Rao, N.K., Pandey, G., & Ivans, I.I., 2001, ApJ, 555, 925

Li, A., & Draine, B.T., 2001, ApJ, 554, 778

Maier, J.P., 1994, Nature, 370, 423

Misawa, T., Gandhi, P., Hida, A., Tamagawa, T., & Yamaguchi, T., 2009, ApJ, 700,1988


Nuccitelli, D., Richter, M.J., & McCall, B.J., 2005, IAU Symp., 235, 236

Pascoli, G., & Polleux, A., 2000, A&A, 359, 799

Scott, A., Duley, W.W., & Pinho, G.P., 1997, ApJ, 489, L193

Sellgren, K., Uchida, K.I., & Werner, M.W., 2007, ApJ, 659, 1338

Smith, J.D.T., et al., 2007a, ApJ, 656, 770

Smith, J.D.T., et al., 2007b, PASP, 119, 1133

Snow, T.P., & Seab, C.G., 1989, A&A, 213, 291

Sogoshi, N., Kato, Y., Wakabayashi, T., et al., 2000, J. Phys. Chem. A., 104, 3733

Somerville, W.B., & Bellis, J.G., 1989, MNRAS, 240, 41P




Yagi, H., et al., 2009, Carbon, 47, 1152

Yasumatsu, H., Kondow, T., Kitagawa, H., Tabayashi, K., & Shobatake, K., 1996, J.Chem. Phys., 104, 899


THE SPITZER SURVEYS OF THE SMALL MAGELLANICCLOUD: INSIGHTS INTO THE LIFE-CYCLE OF POLYCYCLIC

AROMATIC HYDROCARBONS

K.M. Sandstrom1, A.D. Bolatto2 , B.T. Draine3, C. Bot4 and S.Stanimirovic5

Abstract. We present the results of two studies investigating the abun-dance and physical state of polycyclic aromatic hydrocarbons (PAHs)in the Small Magellanic Cloud (SMC). Observations with ISO andSpitzer have shown that PAHs are deficient in low-metallicity galaxies.In particular, galaxies with 12 + log(O/H) < 8 show mid-IR SEDsand spectra consistent with low PAH abundance. The SMC provides aunique opportunity to map the PAH emission in a low-metallicity (12+ log(O/H) ∼ 8) galaxy at high spatial resolution and sensitivity tolearn about the PAH life cycle. Using mid- and far-IR photometry fromthe Spitzer Survey of the SMC (S3MC) and mid-IR spectral mappingfrom the Spitzer Spectroscopic Survey of the SMC (S4MC) we deter-mine the PAH abundance across the galaxy. We find that the SMCPAH abundance is low compared to the Milky Way and variable, withhigh abundance in molecular regions and low abundance in the diffuseISM. From the variations of the mid-IR band strengths, we show thatPAHs in the SMC are smaller and more neutral than their counterpartsin more metal-rich galaxies. Based on the results of these two studieswe propose that PAHs in the SMC are formed with a size distributionshifted towards smaller grains and are therefore easier to destroy undertypical diffuse ISM conditions. The distribution of PAH abundance inthe SMC suggests that PAH formation in molecular clouds is an im-portant process. We discuss the implications of these results for ourunderstanding of the PAH life-cycle both at low-metallicity and in theMilky Way.

1 Max-Planck-Institute for Astronomy, Konigstuhl 17, 69117 Heidelberg, Germany2 Department of Astronomy and Laboratory for Millimeter-Wave Astronomy, University ofMaryland, College Park, MD, USA3 Department of Astrophysical Sciences, Princeton University, NJ, USA4 Observatoire Astronomiques de Strasbourg, Universite Louis Pasteur, 67000 Strasbourg,France5 Department of Astronomy, University of Wisconsin, Madison, WI, USA



1 Introduction

Polycyclic aromatic hydrocarbons (PAHs) are a crucial component of interstellardust. In addition to being responsible for a significant fraction of the total in-frared emission of a galaxy, they can be a dominant source of photoelectrons andparticipate in chemical reactions across a variety of ISM phases. We would like tounderstand how the amount and characteristics of PAHs vary with galaxy type,star-formation history and ISM environment. All of these issues are tied to the lifecycle of PAHs–the balance between their formation, destruction and processing inthe ISM. PAH formation is thought to occur in the atmosphere of evolved stars(Latter 1991; Cherchneff et al. 1992), although it is not clear that evolved starsalone can account for the abundance of PAHs we observe in the ISM (Matsuuraet al. 2009). It may also be the case that PAH formation in the ISM can bea significant source of PAHs (for a recent review, see Draine 2009). PAHs areobserved to be destroyed in ionized gas in H II regions. Shock waves have beenargued to be either a source (via grain shattering; Miville-Deschenes et al. 2002)or sink (Micelotta et al. 2010) of PAHs.

Although many aspects of the life cycle of PAHs are still poorly understood, wedo know that it changes dramatically as a function of metallicity (see contributionsby Calzetti et al. and Hunt et al. in this Volume). Observations with ISO andSpitzer have demonstrated that low metallicity galaxies have a deficit of PAHsrelative to the total amount of dust (e.g. Engelbracht et al. 2005). Draine et al.(2007) used SED modeling to determine the PAH fraction qPAH (defined as thefraction of the dust mass contributed by PAHs with sizes less than 103 carbonatoms) in the SINGS galaxies. They found that at a metallicity of 12 + log(O/H)∼8, galaxies transition from having a median qPAH∼ 3.5%, similar to the Milky Wayvalue of 4.6%, to a median of ∼1%. Spectroscopic (Wu et al. 2006; Engelbrachtet al. 2008) and photometric investigations (Marble et al. 2010) of large samplesof low-metallicity galaxies provide clear evidence for this deficit.

The cause of the low-metallicity PAH deficit has been the subject of muchinvestigation. Galliano et al. (2008) suggest that the deficit is related to PAHproduction due to the time delay between supernova and evolved star enrichmentof the ISM. Others have argued that the deficit is due to destruction of PAHs byharder and more intense UV fields (Madden 2000; Gordon et al. 2008) or enhancedsupernova shock destruction (O’Halloran et al. 2006). The Local Group galaxies,particularly the Small Magellanic Cloud, provide a test case for these scenarios.With a distance of ∼60 kpc and a metallicity of 12 + log(O/H) ∼ 8, the SMC isa location where we can study PAHs at high spatial resolution and sensitivity ina galaxy right at the metallicity of the transition to PAH deficiency.

We have performed two surveys of the SMC using the Spitzer Space Telescope.The first, the Spitzer Survey of the SMC (S3MC), imaged the main star-formingareas of the Wing and Bar of the galaxy using all of the photometric bandsof IRAC and MIPS. Details of S3MC and the data processing can be found inBolatto et al. (2007). The second survey, the Spitzer Spectroscopic Survey of theSMC (S4MC), consisted of extensive spectral mapping observations using the low

K.M. Sandstrom et al.: PAHs in the SMC 217

Fig. 1. Best fit models for photometry alone (photofit) and photometry and spectroscopy

combined (photospectrofit) for a few selected regions covered by S3MC and S4MC.

resolution modules (5–35 μm) of IRS covering six of the main star-forming regionsin the galaxy. The details of the survey and the data reduction can be found inSandstrom et al. (2009) and Sandstrom et al. (2010, in prep.).

2 The PAH fraction in the SMC

In order to understand the cause of the PAH deficit in low-metallicity galaxies,we would like to answer the following questions about the SMC: 1) is the SMCdeficient in PAHs, 2) are there spatial variations in the PAH fraction in the SMC,and 3) if there are spatial variations, with what are they correlated? To answerthese questions we have made a map of the PAH fraction qPAH across the galaxyby fitting the SED models of Draine & Li (2007) at every independent pixel inthe map. Details of the SED modeling are described in Sandstrom et al. (2010)along with a more thorough description of the results. In addition to modelingthe SED, we incorporate the S4MC spectroscopy in the overlap regions to testwhether the photometric determination of qPAH does an adequate job of judg-ing the PAH fraction, with essentially only a measurement of the 8 μm feature.Figure 1 shows the best fit models to the photometry only and combined photome-try and spectroscopy for a few locations in our maps. We find that the photometriconly determination of qPAH is very close to the best fit qPAH from the combinedfit for all of the overlap regions.

In Figure 2 we show the PAH fraction map that results from our SED fits. Wefind that the SMC has an average qPAH= 0.6%, significantly below the Milky Wayvalue of qPAH,MW = 4.6%. There are large spatial variations in the SMC PAH


Fig. 2. The PAH fraction maps that result from our SED fitting overlaid with contours

of CO J= (1−0) from the NANTEN survey (left) and carbon star surface density (right).

It is clear that the PAH fraction is correlated with the locations where molecular gas is

present but not with the distribution of evolved stars.

abundance. At maximum qPAH can be within a factor of two of the Milky Waylevel, but most of the galaxy has qPAH= 0.4%, which represents the lower limitof our models. For its metallicity, the SMC average qPAH agrees well with theaverage qPAH∼ 1% found by Draine et al. (2007) for the SINGS sample.

Next, we compare the PAH fraction to various proposed drivers for the PAHabundance. Evolved stars are thought to be the dominant source of PAHs. Wetrace the distribution of evolved stars using color selection criteria on the 2MASScatalog to select carbon AGB stars (Cioni et al. 2006). The distribution of carbonstars follows the well known spheroidally distributed population of old stars in theSMC (Harris & Zaritsky 2004). There is no correspondence between the evolvedstar density and the PAH fraction, in contrast to what is seen in the LMC, wherethe PAH fraction is seen to be enhanced along the stellar bar (Paradis et al. 2009).We find that the PAH fraction is correlated with the presence of molecular gas astraced by CO J = (1–0) from the NANTEN survey (Mizuno et al. 2001). Sincethe molecular clouds condense out of the diffuse ISM on timescales of ∼25 Myr(Fukui et al. 1999), there must be some process which allows the molecular cloudsand the diffuse ISM they condensed out of to have different PAH fractions. Twopossibilities are: 1) a recent (<25 Myr) event which destroyed PAHs in the dif-fuse ISM over the whole galaxy after the current generation of molecular cloudscondensed; or 2) some component of interstellar PAH material forms in molecularclouds. The SMC star-formation history shows little evidence for a recent eventthat could have altered the PAH fraction globally, so we consider the latter optionmore likely.


3 The physical state of PAHs in the SMC

The ratios of the mid-IR PAH bands provide a wealth of information about thephysical state of PAHs–charge, size, structure and composition all alter the relativestrengths of the mid-IR features. Understanding the physical state of PAHs inthe SMC may help us to explain the low-metallicity PAH deficit. For instance,PAHs that are smaller or ionized are easier to photodissociate under typical ISMconditions (Allain et al. 1996a, 1996b). To study the physical state of PAHs inthe SMC, we have used the S4MC spectroscopy and measured the strengths ofthe various bands. A detailed description of this work can be found in Sandstromet al. (2010, in prep.).

In Figure 3 we show measurements of the band ratios overlaid on theoreticalpredictions from Li & Draine (2001). The gray lines show neutral (top) andsingly-ionized (bottom) PAHs with a variety of sizes (labeled on the bottom plot).Since interstellar PAHs come with a size and charge distribution, the theoreticalpredictions are only used as a guide for interpreting the ratios relative to othermeasurements. We also plot the band ratios measured in the SINGS galaxies(without AGN; Smith et al. 2007). The SINGS spectroscopy covers regions withapproximately solar metallicity. Compared to SINGS, we find that the SMC PAHstend to be more neutral and smaller, judging from the ratios of their 6 − 12 μmfeatures. In addition, we find that the 17 μm PAH complex is notably weak in theSMC, continuing a trend noted by Smith et al. (2007) for the strength of the 17feature to decrease as a function of metallicity.

4 Insights into the PAH life cycle

To summarize, the Spitzer surveys of the SMC (S3MC and S4MC) allow us to studythe abundance and physical state of PAHs in a nearby, low-metallicity galaxy. Wefind that the SMC PAH fraction is significantly below the Milky Way value, butwith large spatial variations. High PAH fractions are found in molecular regions,while the diffuse ISM is essentially devoid of PAHs (to the limit of our models).Spectroscopic investigations of a number of the main star-forming regions showthat the mid-IR band ratios suggest smaller and more neutral PAHs than foundin more metal rich galaxies (for example in the SINGS sample).

These results suggest a scenario where PAH formation in dense regions is im-portant, at least at low-metallicity. The process by which this proceeds is currentlyuncharacterized, but may depend on the coagulation of material on the surfacesof existing grains. The PAHs that result from this formation channel in the SMCseem to be smaller on average than PAHs in more metal rich galaxies, judging bytheir band ratios. Smaller PAHs are easier to photodissociate under typical diffuseISM conditions, perhaps leading to the very low PAH abundance over most of theSMC. PAHs formed in the atmospheres of evolved stars are mainly injected firstinto the diffuse ISM and may be largely destroyed in the SMC. At higher metal-licity, we speculate that PAHs are better able survive in the diffuse ISM, possiblybecause of softer and less intense UV fields and/or larger average PAH sizes. As


Fig. 3. Ratios of the 6.2, 7.7 and 11.3 μm PAH bands in the SMC compared with the

SINGS sample and the theoretical models of Li & Draine 2001. The band ratios suggest

that SMC PAHs are smaller and more neutral than PAHs in the SINGS galaxies.

PAHs become able to survive in the diffuse ISM, the overall PAH fraction of agalaxy can build up to Milky Way levels. Resolved studies of the PAH fraction ina larger sample of nearby galaxies that span a range of metallicity will allow us tostudy these trends.

References

Allain, T., Leach, S., & Sedlmayr, E., 1996a, A&A, 305, 602

Allain, T., Leach, S., & Sedlmayr, E., 1996b, A&A, 305, 616

Bolatto, A.D., Simon, J.D., Stanimirovic, S., et al., 2007, ApJ, 655, 212


Cioni, M.-R.L., Girardi, L., Marigo, P., & Habing, H.J., 2006, A&A, 452, 195

Draine, B.T., 2009, in Astronomical Society of the Pacific Conf. Ser., 414, ed. T. Henning,E. Grun, & J. Steinacker, 453

Draine, B.T., & Li, A. 2007, ApJ, 657, 810


Engelbracht, C.W., Gordon, K.D., Rieke, G.H., et al., 2005, ApJ, 628, L29



Fukui, Y., Mizuno, N., Yamaguchi, R., et al., 1999, PASJ, 51, 745



Harris, J., & Zaritsky, D., 2004, AJ, 127, 1531

Latter, W.B., 1991, ApJ, 377, 187

Li, A., & Draine, B.T., 2001, ApJ, 554, 778

Madden, S.C., 2000, New Astron. Rev., 44, 249

Marble, A.R., Engelbracht, C.W., van Zee, L., et al., 2010, ApJ, 715, 506

Matsuura, M., Barlow, M.J., Zijlstra, A.A., et al., 2009, MNRAS, 396, 918

Micelotta, E.R., Jones, A.P., & Tielens, A.G.G.M., 2010, A&A, 510, A36

Miville-Deschenes, M.-A., Boulanger, F., Joncas, G., & Falgarone, E., 2002, A&A, 381,209

Mizuno, N., Rubio, M., Mizuno, A., et al., 2001, PASJ, 53, L45


Paradis, D., Reach, W.T., Bernard, J.-P., et al., 2009, AJ, 138, 196


Sandstrom, K.M., Bolatto, A.D., Stanimirovic, S., van Loon, J.T., & Smith, J.D.T.,2009, ApJ, 696, 2138


Wu, Y., Charmandaris, V., Hao, L., et al., 2006, ApJ, 639, 157


PAH-RELATED VERY SMALL GRAINSIN PHOTODISSOCIATION REGIONS: IMPLICATIONS

FROM MOLECULAR SIMULATIONS

M. Rapacioli1, F. Spiegelman 1, B. Joalland1, 2, A. Simon1,A. Mirtschink1, C. Joblin2, J. Montillaud2, O. Berne3 and D. Talbi4

Abstract. The analysis of mid-IR emission suggests that a populationof PAH-related very small grains containing a few hundreds of atomsare present in the deep regions of molecular clouds, although no specificspecies has been identified yet. In this review, we discuss several candi-dates for these grains: neutral and ionised PAH clusters and complexesof PAHs with Si atoms. The theoretical modelling of the propertiesof such molecular complexes or nanograins is a challenging task. Wefirst present an overview of quantum chemistry derived models whichcan be efficiently used on-the-fly in extensive sampling of the poten-tial energy surfaces, as required by structural optimization, classicalmolecular dynamics or Monte Carlo algorithms. From the simulations,various observables can be determined, such as the binding energies,finite temperature IR spectra, nucleation and evaporation rates. Wediscuss the relevance of those candidates in the molecular clouds pho-todissociation regions and propose constrains and perspectives for thenature and size of those very small grains.

1 Introduction

Polycyclic Aromatic Hydrocarbons (PAHs) are usually considered to be at the ori-gin of the Aromatic Infrared Bands (AIBs) observed in emission from

1 Laboratoire de Chimie et Physique Quantiques, Universite de Toulouse (UPS) and CNRS,IRSAMC, 118 route de Narbonne, 31062 Toulouse, France2 Universite de Toulouse, UPS, CESR, 9 avenue du colonel Roche, 31062 ToulouseCedex 4, France and CNRS, UMR 5187, 31028 Toulouse, France3 Leiden Observatory, Leiden University, Niels Bohrweg 2, 2333 CA Leiden,The Netherlands4 Universite Montpellier II - GRAAL, CNRS - UMR 5024, Place Eugene Bataillon,34095 Montpellier, France



UV-irradiated interstellar matter (Allamandola et al. 1985; Leger & Puget 1984,see also Peeters, this volume). Boulanger et al. (1990), Bernard et al. (1993) anal-ysed the emission in the IRAS photometric bands measured for several molecularclouds, and suggested that free-flying PAHs are produced by photoevaporationof larger grains. The observation of a mid-IR continuum in the reflection nebulaCed 201 has been attributed by Cesarsky et al. (2000) to Very Small carbonaceousGrains (VSGs) that have been proposed as precursors of the AIB carriers.

Blind signal separation methods like non-negative matrix factorisation or sin-gular value decomposition (combined with a Monte Carlo search of physical so-lutions) have been applied to extract three characteristic spectra from mid-IRspectro-imagery data (ISO and Spitzer) of the PhotoDissociation Region (PDR)NGC 7023 (Berne et al. 2007; Rapacioli et al. 2005, see Berne, this volume). Twospectra have been assigned to PAH populations dominated either by neutral orionised molecules. The third spectrum is composed of broader AIBs associatedwith a continuum. As these two characteristic features suggest larger systems,this component has been attributed to a population of PAH-related VSGs. Thespatial distribution of these spectra shows that the ionised PAH population is lo-cated around the star followed by the neutral PAH population and then by theVSG population when going further from the star and inside the molecular cloud.This suggests a chemical scenario in which PAHs are trapped in VSGs that arephotoevaporated at the border of the cloud where the UV flux increases. Closerto the star, the UV flux is strong enough to keep the PAH population mainlypositively ionised.

The presence of VSGs related to PAHs has been reported in several PDRs(NGC 7023, ρ Ophiuchi, Ced 201; Berne et al. 2007) but also in protoplanetarydisks (Berne et al. 2009) and in planetary nebulae both in the Milky Way and inthe Magellanic clouds (Joblin et al. 2008).

A minimal size of these grains can be estimated to 400 carbon atoms fromthe grain heating process as described in Rapacioli et al. (2006). In this paper,we review different hypotheses for PAH-related VSGs and show the capability oftheoretical modelling as a tool to investigate these clusters. Three possible speciesare mostly discussed here: neutral and ionised PAH clusters and complexes ofPAHs with Si atoms.

2 Possible candidates for the PAH-related VSGs

Several hypothesis can be made for the PAH-related VSGs. Possible candidatesthat will be discussed in the following are:1) Neutral PAH clusters: based on the sequential evolution PAH+,PAH0, VSGthat is observed in molecular clouds, the most intuitive candidates are PAH clus-ters. As these clusters are located in deeper regions of molecular clouds than theneutral PAH population, it seems consistent to investigate neutral PAH clusters.2) Ionised PAH clusters: as will be discussed in Section 5, the lowering of the ioni-sation potential (IP) in PAH clusters compared to PAH could allow the presence ofionised PAH clusters deeper inside molecular clouds. This hypothesis is reinforced

M. Rapacioli et al.: PAH-related Very Small Grains 225

because -(i)- recent calculations (Rhee et al. 2007) show that small closed-shellPAH cationic dimers could be at the origin of the Extended Red Emission (ERE)and -(ii)- in NGC 7023N, the ERE spatially correlates with the maximum of VSGemission (Berne et al. 2008).3) Complexes of PAHs with an heteroatom: complexes with Fe atoms can formeasily from energetics considerations and have recently been studied both experi-mentally and theoretically (Simon & Joblin 2007, 2009, 2010; Simon et al. 2008)and are the subject of a specific contribution (Simon et al.) in this volume. Ina similar way, complexes of PAH molecules or clusters with Si atoms could beconsidered as candidates for the VSGs and are presented in the last section.This list is of course not exhaustive and the VSGs are most probably made ofmixture of various types of grains rather than belonging entirely to one of thepreviously enumerated families.

3 Theoretical methods adapted to large systems

In the following, we present specific methods that we have used with some the-oretical development when needed. The treatment of large clusters containinga few hundreds of atoms is a challenging task for theoretical chemistry. TheDensity Functional Theory (DFT) is the most widely used method to describeelectronic structure, based on a density spanned on molecular orbitals and the so-lution of a mean-field type single electron Kohn-Sham equation. However, despitefast algorithmic developments, it can presently hardly treat such large clusters, inparticular as soon as global Potential Energy Surface (PES) exploration or molec-ular dynamics (MD) are needed. An alternative for even larger systems and/orlarger computational efficiency is provided by approximate schemes of DFT, forinstance the Density Functional based Tight-Binding approach (DFTB, Elstneret al. 1998; Oliveira et al. 2009; Porezag et al. 1995; Seifert et al. 1996). DFTB isbased on a second order Taylor expansion of the DFT energy expression arounda reference electronic density and the expression of the second order term as afunction of atomic charges. The molecular orbitals are spanned in a minimal setof valence atomic orbitals. All integrals are parametrised and the ground state isfound by solving self-consistently the approximate Kohn-Sham secular equation(Self-Consistent-Charge DFTB), including electrostatics self-consistently. In prin-ciple, DFTB parametrisation is extracted from DFT calculations on the variousatom-pairs and may be expected to show satisfactory transferability. Nonetheless,several parametrisations do exist, adapted to various situations (bulk, biologi-cal systems). Parametrisation adjustment may be sometimes needed. In furtherapproximations, the self-consistency can be circumvented, or the electrostatics to-tally ignored, leading to simple Tight-Binding (TB) schemes. At the present time,DFT (with the most common functionals) and DFTB experience theoretical dif-ficulties such as (i) their failure to describe long-range dispersion forces; (ii) thefailure to describe dissociation properly partly screened in the spin-unrestrictedformulation; (iii) the difficulty to describe properly open-shell system and mul-tiplets; (iv) the difficulty of todays functionals in the Kohn-Sham formalism to


properly describe systems with strong multi-configurational character. This hasmotivated the search for new functionals and also ab initio hybrid methods basedon both the wavefunction for long-range electron-electron correlation and DFTfor the short-range correlation (Chabbal et al. 2010; Fromager & Jensen 2008;Toulouse et al. 2004). The lack of dispersion can also be corrected by additionof empirical atom-atom dispersion potentials in both DFT (Grimme 2006) andDFTB schemes (Elstner et al. 2001; Rapacioli et al. 2009; Zhechkov et al. 2005).

Moreover, in the case of molecular clusters, one can benefit from the partitionof the cluster into identifiable units to propose further modelling with differentlevels of treatments for intra- and intermolecular degrees of freedom. This givesrises to a partition of the level of treatment of the degrees of freedom such asfor instance in the case of neutrals: (i) a hybrid approach in which the molecularunits (not rigid) are treated via DFT, DFTB or TB (actually developed, see below)and the intermolecular interactions (repulsion, electrostatics and dispersion) aretreated via atom-atom classical potentials. This provides a consistent descriptionbetween intra-molecular charge variations and intermolecular electrostatics. Suchmodel can be considered as a special version of quantum-mechanics/molecular(classical) mechanics (QM-MM, i.e. DFTB-MM) schemes. (ii) a more basic rigidmolecule approximation (for a review see Stone 1997), where the molecules con-stituting a cluster are treated as rigid frozen systems and only the intermolecularpotential is considered. The rigid body approach obviously is the fastest one andallows global exploration of the intermolecular PES for quite large systems withhundreds of atoms. The advantage of the hybrid (DF)TB/MM approach is itsability to describe energy transfers during collisions or couplings between inter-and intramolecular vibrational modes.

The molecular partition is also relevant for resonance in ionised molecularclusters. Specific approaches consist in expressing the multiconfigurational wave-function of the ionized cluster as an expansion of charge-localised configurations(Wu et al. 2007), A small Configuration Interaction matrix (CI) is then built inthis limited valence-bond-like multiconfigurational basis. Such DFT-CI model isclearly very well suited to model charged molecular clusters. It can even be mademore efficient in the scope of large systems when using DFTB instead of DFT:benchmark calculations on the benzene, coronene and water dimers show that theDFTB-CI method (Rapacioli et al. 2010) is able to correct the wrong behaviorof the DFTB dissociation pathway and gives binding energies in agreement withexperimental data and reference calculations.

4 Neutral PAH clusters

4.1 Structural properties

4.1.1 PAH dimers

The intermolecular interactions vanish quickly with increasing distance, except forthe electrostatic terms which essentially account for interactions between permanent


Fig. 1. Stacked coronene clusters of 4 and 10 molecules.

multipoles. The structural growth patterns of PAH clusters are therefore closelyrelated to the dimer structures. Stacks are expected if the most stable dimerstructures have sandwich-like configurations. The structures of PAH dimers rang-ing from benzene to coronene have been optimized at different levels of theory inrecent years. The structure of the benzene dimer has been a long-standing dis-cussion until high-level calculations (CCSD(T), SAPT see for instance Podeszwaet al. 2006) showed that the PES of this dimer has several almost degeneratedstructures corresponding to T-shaped and sandwich-parallel-displaced geometries.The stacked sandwiches are less stable. For larger PAHs, the T-shaped config-urations are strongly destabilised as compared to sandwich structures both inparallel-displaced and stacked configurations. The binding energy increases fromabout 0.12 eV for the benzene dimer to 0.8-1eV for the coronene dimer (Rapacioliet al. 2009; Zhao & Truhlar 2008). The rigid body, TB-MM and DFTB approacheshave been parametrised and benchmarked (Rapacioli et al. 2005, 2006, 2009) onthese high-level calculations on PAH dimers.

4.1.2 Large coronene clusters

A global exploration of the PES using parallel tempering Monte-Carlo methodswith the rigid body approach provides the structures of coronene clusters con-taining up to 32 molecules (Rapacioli et al. 2005). As expected from the dimersandwich structures, the lowest-energy structures of coronene clusters are foundto be single stacks for numbers of molecules not exceeding 8. For larger clusters,lower energy structures appear made of two stacks with either parallel axis or ina handshake-like structure (axes almost perpendicular, Fig. 1). The special sta-bility of the “handshake” structure is due to its compactness, while conservingthe primary stack structure as basic units. The (PAH)32 cluster presents a largenumber of degenerated isomers which all consist of small units of 4 molecule stacksaggregated to each others. The dissociation energy of these clusters (energy re-quired to remove one molecule) is almost constant (about 1 eV) for clusters above3 molecules as the less bounded molecules are always at the tip of one stack andinteract almost only with their first two neighbors.

In the interstellar medium, PAH clusters are not expected to be made of ho-mogenous units but to aggregate PAHs of different sizes whose structures can alsobe calculated. The stacks in which the larger PAHs occupy inner sites maximize thenumber of nearest neighbors between carbon atoms and are energetically favored.


This segregation property remains for larger clusters made of several stacks, thesmallest molecules being located at the tips of the stacks.

4.2 Stability in NGC 7023N

A key question for astrophysics is whether these clusters can survive in regionssuch as NGC 7023N where they have been proposed to be present. Their stabilityresults from a competition between formation, upon PAHs/clusters collisions, anddestruction by the UV flux. The corresponding rates have been estimated frommolecular modelling simulations for clusters made of 4 or 13 coronene molecules(Rapacioli et al. 2006). Formation rates were calculated from collision rates be-tween molecules and clusters, which depend on the temperature and density ofthe cloud, and the probability of sticking at each collision. This last value iscalculated from MD simulations using the TB-MM approach combining the TBmodel of Van Oanh et al. (2002) with a classical intermolecular pair potentials asmentioned above. To describe the evaporation which might follow a UV photonabsorption, one needs to compare the rate of energy relaxation either by IR emis-sion or by evaporation of one molecule. Both phenomena can occur on relativelylong timescales (at low energy) and real-time simulations of these processes arenot possible. However these values can be obtained with statistical approaches,namely the Phase Space Theory for the molecular evaporation (Calvo & Parneix2004) with the rigid body model and a microcanonical IR emission model (Joblinet al. 2002) for the IR cooling.

The calculated clustering timescales are typically 104–8×104 years, much longerthan the photoevaporation timescales which are below 17 years. Therefore, theseclusters should be destroyed by the UV flux much faster than they can be reformed(Rapacioli et al. 2006). Such clusters could not survive in UV irradiated cloudslike the PDR NGC 7023N. These results do not however rule out completelythe hypothesis of neutral PAH clusters as the size of astronomical PAH-relatedVSGs might be larger (larger clusters or larger units) and this would increasetheir stability in PDRs.

5 Ionised PAH clusters

Ionised molecular clusters are more strongly bound than the neutrals. The chargeresonance at the origin of this strong stabilisation, completed by polarization, istreated with the DFTB-CI method.

5.1 Energetics considerations

We have investigated ionised coronene clusters in non-relaxed stacked geometries(Rapacioli & Spiegelman 2009). The charge essentially delocalises over the centralmolecules: the configurations where the charge is carried by the central moleculeshave much higher weights in the global CI wavefunction than the configurationswhere the charge is localised on the edge molecules.


2 4 6 8Number of units

16

18

20

22

24

26

28

30

Dis

soci

atio

n E

nerg

y (k

cal/m

ol)

Neutral clusterIonised cluster

Fig. 2. Dissociation energy of neutral and ionised stacked coronene clusters (Fig. from

Rapacioli & Spiegelman 2009).

Figure 2 displays the dissociation energy of neutral and cationic coronene stackscorresponding to the evaporation of a single molecule from one stack side. Inthe neutral cluster, the dissociation energy reaches a limit for sizes above 3 unitsas seen previously. The cationic dimer is over 1.5 times more stable than theneutral one. The evaporation rate constant varies approximately exponentially asa function of the dissociation energy, thus cationic clusters will be significantlymore resistant to the loss of a neutral molecule. When cluster size increases, thecharge distribution is less sensitive to the removal of one unit and the dissociationenergy tends to the same limit as observed for neutral clusters.

It should also be noticed that, as the charge delocalisation over the stack sta-bilises the ionised form, the IP decreases when the cluster size increases. Thedecrease of the IP with cluster size down to the bulk limit is frequently observedfor various families of clusters. For a coronene stack, the IP reaches a limit forclusters above 7–8 units. The IP is then reduced by more than 1 eV compared tothat of the isolated molecule. This decrease of the IP could explain the presenceof cationic PAH clusters in PDRs regions where isolated PAHs would be neutral.Most of the features will certainly remain valid when achieving full relaxationof the ion geometries. The issues are (i) how large is the core subcluster overwhich the charge is actually delocalised; (ii) what is the quantitative interplaybetween geometric relaxation and stability, and the construction of larger clusters;(iii) what is the supramolecular organization of the other units essentially neutraland polarized by this core; and (iv) what are the implications for nucleation andevaporation. Implementation of the analytical gradient in DFTB-CI is underwayand should allow to answer precisely these questions.

5.2 Formation rates

Although the complete study of ionised PAH cluster formation would require toperform MD to simulate collisions as done for neutral clusters, the evolution of thecollision rate can be estimated. Because the VSGs are observed in regions wherePAHs are expected to be neutral, it is consistent to focus on collisions between aneutral molecule with a (PAH)+n cluster. In such a collision, the cationic cluster


induces a dipole on the PAH molecule. The long-range ionic clusters/induceddipole interaction is attractive and scales as 1/R4. This interaction is muchstronger than in the case of neutral clusters, which scales as 1/R6. The formationrate can be roughly estimated from a simple Langevin model. The Langevin pre-diction for the nucleation rate is then given by knucl = 2e(α/μ)1/2 (see for instanceHerbst 2001), where e is the elementary charge, α the polarisability of the PAH,and μ the reduced mass of the reactants. This leads, for the specific case of NGC7023N, to reformation timescales of a few thousands years, much faster than forneutral clusters. The conclusion that small clusters are destroyed faster than theycan be reformed could thus be modified for ionised clusters, but a complete studyof their stability, including the calculation of the evaporation rates, necessary toanswer this question, is under development.

6 SiPAH complexes

PAH-related VSGs could contain heteroatoms. The case of Fe is discussed inSimon et al. (this volume). We discuss here SiPAH complexes. The first step inthe study of such systems consists in investigating the smallest dimers, that is asingle PAH molecule with a single Si atom. Some characteristic features associatedto the coordination of Si can already be derived from these simple model systems,in particular concerning the structures and IR spectra of [SiPAH ]+ complexes.

6.1 Structures and binding energies

The structures of [SiPAH]0/+ complexes have been optimized at the DFT levelby Joalland et al. (2009). The SiPAH binding energies in [SiPAH]+ complexeswere computed to be at least 1.5 eV. This value being three times higher thanfor [SiPAH ]0, [SiPAH]+ complexes are therefore more likely to survive in theconditions of the ISM than their neutral counterparts. Furthermore, the formationof [SiPAH]+ complexes by radiative association is an exothermic process withoutany expected activation barrier. These [SiPAH]+ complexes could then representa significant fraction of PAH-related species and their presence has been suggestedto account for the blue-shift of the 6.2 μm band (see also Peeter et al., this volume).The exploration of the PES reveals that in the most stable structures, the Si atomis located on top of a C atoms on the edge of the PAH molecular structure. Itwas furthermore demonstrated that the lowest energy path for the motion of theSi atom parallel to the PAH surface is through the edge of the PAH.

6.2 IR spectroscopy

6.2.1 Harmonic

The IR spectra of such complexes have been calculated with the harmonic approx-imation at the DFT level. Compared to isolated PAHs, the [SiPAH]+ complexespresent very soft modes associated to the motion of the Si atom above the PAHsurface. For instance, in [SiC10H8]+ complex, these modes are located at 202, 240,


and 413 cm−1. The first two correspond to a Si motion parallel to the PAH planeand have very low intensities. The 413 cm−1 mode corresponds to a Si motion ina direction perpendicular to the PAH plane with an intensity of about 12% of themaximum intensity.

The mid-IR spectra of [SiPAH]+ resemble more that of a cationic PAH thanthat of a neutral PAH. This results from the fact that in the [SiPAH]+ complexes,most of the charge is actually carried by the PAH. A significant spectral differencestems from the fact that the symmetry is reduced in the complex. For instance,the γCH band located at 780 cm−1 in C10H

+8 splits into two subbands at 862

and 786 cm−1 in [SiC10H8]+. The first band involves mainly the bending of C-Hbonds from the ring neighboring the Si atom whereas the second γCH band involvesthe H atoms coordinated to the other rings and is therefore less perturbed. The1594 cm−1 band is assigned to a νCC mode involving the whole PAH skeleton. It isblue-shifted by 29 cm−1 with respect to that in C10H

+8 . The splitting of the γCH

band into two bands and the blue-shift of the νCC band are therefore signaturesto probe the presence of [SiPAH]+complexes in the ISM.

6.2.2 Anharmonic

The spectra calculated at the DFT level are harmonic spectra corresponding tosmall amplitude motions (spectra at the limit of T = 0 K). In the emission pro-cess observed in the ISM, the temperature of such complexes differs from 0 Kand anharmonic effects should be taken into account. In the classical motionapproximation, these effects can be derived from MD simulations by computingthe Fourier transform of the complex dipole autocorrelation function. The use ofDFTB on-the fly allows to perform MD simulations on the required long-timescales(Joalland et al. 2010) and to explore various ranges of temperatures.

For low temperatures, the Si atom remains on top of the same ring. For highertemperatures, there is sufficient energy for the Si atom to overcome the barrierseparating two PES minima (differing by the Si position on top of two adjacentcycles) and the average time spent on a given ring decreases as the temperatureincreases.

Figure 3 shows the evolution of the γCH mode in the naphthalene cation. Itexhibits a broadening and a red-shift when the temperature increases. The samebehavior is observed in the [SiC10H8]+ complex for the two γCH bands that resultfrom the splitting due to the Si coordination. However, the broadening is muchmore significant for temperatures larger than 600 K, when the system reaches thehigh temperature mobility regime. At this temperature, the Si oscillation over thetwo rings is fast enough to couple the two C-H modes. The two correspondingbands are thus shifted towards each other, each one displaying a broad wing inthe interband region.

The study of complexes with increasing sizes shows that this band broadeningcan be generalized to the spectra of [SiPAH]+ complexes compared to those of thecorresponding bare PAH+. Although the spectra calculated for a limited number


11

11.5

12

12.5

13

13.5

100

200

300

400

500

600

700

00.5

1

Wavelength (μm)Temperature (K)

Inte

nsity

(ar

b. u

n.)

Fig. 3. Evolution of the γCH band in bare C10H+8 (blue/open circles) and in the

[SiC10H8]+ complex (red/filled circles) as a function of the temperature.

of SiPAH complexes are not directly able to account the VSG spectrum, this bandbroadening is consistent with the features observed in the VSG spectrum.

7 Conclusion

The spectral evolution of the AIBs in PDRs provides evidence for PAH evolutionin molecular clouds. Inside molecular clouds, the spectra present characteristics ofsmall carbonaceous grains which seem to evaporate, where the UV flux becomesstrong enough, to give the free-flying PAHs. The theoretical modelling of thesegrains requires multiscale approaches combining different levels of theory to modelthe PES.

Possible candidates for these grains are PAH clusters in their neutral form asthey are located in regions where PAHs are expected to be neutrals. A globalexploration of the structural properties shows that the most stable structures are


self assemblies of small units made of PAH stacks. The comparison of calcu-lated nucleation rates for clusters of 4 and 13 coronene molecules versus theirevaporation rates in the physical conditions of NGC 7023N leads to the conclu-sion that such clusters are photoevaporated faster than they can be reformed.They are thus not the observed VSGs and other hypotheses must be investigatedlike for instance clusters made of larger PAHs and/or with more units. Othercandidates are cationic PAH clusters. The binding energies are larger than thatcomputed for their neutral counterparts and the long-range attraction of the ioniccluster/polarisable molecule strongly increases the formation rate. Finally we in-vestigated the possibility that VSGs could consist of complexes of PAHs withheteroatoms such as Fe or Si. The harmonic spectra of [SiPAH]+ complexes showthe appearance of new bands compared to the spectra of bare PAH+. When thetemperature increases, these bands are broadened, shifted and mixed to give broadstructures. Such effects are in line with the broadening of the AIBs in the VSGspectra.

The question of the VSG nature is still an open question. At the moment, theonly conclusion that can be drawn is that these grains are larger than PAHs butstrongly related to them. If the small neutral PAH clusters hypothesis has beenruled out, candidates like clusters of larger molecules and/or with more units,ionised PAH clusters and Fe/Si-PAH complexes need further investigations. Aglobal evolution model of PAHs and related species would be very useful to vali-date/invalidate the various hypotheses (See for instance Montillaud et al., in thisvolume).

The authors thank F. Calvo and P. Parneix for their contributions in various aspects of themodelling and simulations presented in this overview.

References



Berne, O., Joblin, C., Fuente, A., & Menard, F., 2009, A&A, 495, 827

Berne, O., Joblin, C., Rapacioli, M., et al., 2008, A&A, 479, L41

Calvo, F., & Parneix, P., 2004, JCP, 120, 2780

Chabbal, S., Stoll, H., Werner, H.-J., & Leininger, T., 2010, Mol. Phys., submitted

Elstner, M., Hobza, P., Frauenheim, T., Suhai, S., & Kaxiras, E., 2001, JCP, 114, 5149

Elstner, M., Porezag, D., Jungnickel, G., et al., 1998, Phys. Rev.B, 58, 7260

Fromager, E., & Jensen, H.J.A., 2008, Phys. Rev.A, 78, 022504

Grimme, S., 2006, J. Comp. Chem., 27, 1787

Herbst, E., 2001, Chem. Soc. Rev., 30, 168

Joalland, B., Rapacioli, M., Simon, A., et al., 2010, JPC A, 114, 5846

Joalland, B., Simon, A., Marsden, C. J., & Joblin, C., 2009, A&A, 494, 969


Joblin, C., Toublanc, D., Boissel, P., & Tielens, A., 2002, Mol. Phys., 22, 3595



Oliveira, A., Seifert, G., Heine, T., & duarte, H., 2009, J. Braz. Chem. Soc., 20, 1193

Podeszwa, R., Bukowski, R., & Szalewicz, K., 2006, JPC A, 110, 10345

Porezag, D., Frauenheim, T., Kohler, T., Seifert, G., & Kaschner, R., 1995, Phys. Rev.B,51, 12947

Rapacioli, M., Calvo, F., Joblin, C., et al., 2006, A&A, 460, 519

Rapacioli, M., Calvo, F., Spiegelman, F., Joblin, C., & Wales, D., 2005, JPC A, 109,2487


Rapacioli, M., & Spiegelman, F., 2009, EpJD, 52, 55

Rapacioli, M., Spiegelman, F., Mirtschink, A., & Scemama, A., 2010, in preparation

Rapacioli, M., Spiegelman, F., Talbi, D., et al., 2009, JCP, 130, 244304

Rhee, Y.M., Lee, T.J., Gudipati, M.S., Allamandola, L.J., & Head-Gordon, M., 2007,PNAS, 104, 5274

Seifert, G., Porezag, D., & Frauenheim, T., 1996, IJQM, 58, 185

Simon, A., & Joblin, C., 2007, JPC A, 111, 9745

Simon, A., & Joblin, C., 2009, JPC A, 113, 4878


Simon, A., Joblin, C., Polfer, N., & Oomens, J., 2008, JPC A, 112, 8551

Stone, A., 1997, The Theory of Intermolecular Forces, 0198558848, (USA, Oxford:University Press)

Toulouse, J., Colonna, F., & Savin, A., 2004, Phys. Rev. A, 70, 6

Van Oanh, N.T., Parneix, P., & Brechignac, P., 2002, JPC A., 106, 10144

Wu, Q., Cheng, S.-L., & Van Voorhis, T., 2007, JCP, 127, 164119

Zhao, Y., & Truhlar, D. G., 2008, JPC C, 112, 4061

Zhechkov, L., Heine, T., Patchovskii, S., Seifert, G., & Duarte, H., 2005, JCTC, 1, 841


THE FORMATION OF BENZENE IN DENSEENVIRONMENTS

P.M. Woods1

Abstract. PAHs are traditionally thought to form around carbon-richAGB stars, where the presence of inner-wind shocks and nucleationseeds can aid their formation. However, my recent work has shownthat the formation of benzene – thought to be the rate-limiting step inthe formation of larger PAHs – can be efficient in other environments,namely the dense tori around the evolved stars of pre-planetary nebulaeand in the inner regions of the accretion discs around young stars. I willdiscuss the chemical pathways to the formation of benzene in theseregions and the subsequent formation of larger PAHs.

1 Introduction

The first discovery of benzene outside of the Solar System was made by Cernicharoet al. (2001a) towards the carbon-rich pre-planetary nebula2, CRL 618. A smallabsorption at ∼14.8μm in the mid-infrared spectrum from the Short-WavelengthSpectrometer on board the Infrared Space Observatory was attributed to theQ-branch of the ν4 band of benzene. These authors also detected lines of diacety-lene (C4H2) and triacetylene (C6H2), and in the companion paper(Cernicharo et al. 2001b) detected methylpolyynes and other small hydrocarbonstowards CRL 618. The detection of benzene was exciting for several reasons:firstly, this was another molecule to add to the list of 100 or so interstellar andcircumstellar molecules detected at that time (the list has since grown to ∼160).Secondly, benzene (C6H6) is a large molecule, and very few molecules of this sizehave been detected in the interstellar or circumstellar media. Thirdly, benzene is acyclic molecule, more circular in structure than the “triangular” cyclic molecules,

1 Jodrell Bank Centre for Astrophysics, School of Physics & Astronomy, University ofManchester, Oxford Road, Manchester M13 9PL, UKhttp://www.jb.man.ac.uk/~pwoods

2I use the term “pre-planetary nebula” as opposed to “protoplanetary nebula” (both termsdescribing evolved, post-Asymptotic Giant Branch stars) in order to save confusion with “pro-toplanetary disc”, which is the accretion disc around a young, low-mass star.



ethylene oxide (c-C2H4O) and cyclopropenylidene (c-C3H2). As such, its detectionheralded the beginnings of an astrochemistry (and an astrobiology) which was notsolely dependent on the large linear carbon chain molecules previously seen in thearchetypical carbon star, IRC+10216, for example.

Five years subsequently, Bernard-Salas et al. (2006) reported the detection ofbenzene, diacetylene and triacetylene in another galaxy – the Large MagellanicCloud. In this case, again, the object in question was a pre-planetary nebula,SMP LMC 11. Herpin & Cernicharo (2000) suggested, for the case of CRL 618,that the molecular richness originates in a circumstellar torus which is sufficientlydense to shield molecules from the intense UV fields in the environments aroundthese proto-White Dwarf stars. Certainly without sufficient shielding complexmolecules in the post-AGB phase of evolution are rapidly destroyed (Woods et al.2005). In these proceedings I will summarise work which shows that benzene canform efficiently in the post-AGB phase, in excellent agreement with the observa-tions of Cernicharo et al. (2001a). I will also discuss the formation of benzene in aphysically and chemically similar environment, the inner regions of protoplanetarydiscs, where benzene has yet to be detected but PAHs have been seen (Geers et al.2006).

2 Combustion chemistry or interstellar chemistry?

Benzene and polycyclic aromatic hydrocarbons (PAHs) are readily found on Earthin crude oil and coal, and their combustion properties have been well-studied in thelaboratory, motivated in part by the petroleum and combustion engine industriesand latterly environmental concerns. Both benzene and PAHs are known to form inhydrocarbon flames under terrestrial conditions (e.g., Frenklach 2002), and couldpotentially form in circumstellar conditions where the chemistry and physics aresimilar. Such an approach has been used by combustion chemists (e.g., Frenklach& Feigelson 1989) and astrochemists (e.g., Cau 2002; Cherchneff et al. 1992, seealso Cherchneff, this volume) to explain carbon dust (soot) formation in the innerwinds of carbon-rich red giant stars. However, the temperatures reached in flamesand in the inner regions of AGB stars (1000–1500K) are not readily found in theinterstellar and much of the circumstellar media. Thus we must look to otherexplanations for the formation of benzene in pre-planetary nebulae, and given thechemical richness that some of these objects display, it seems logical to considerthe “bottom-up” route to the formation of benzene and PAHs. (An alternativewould be the “top-down” route, where graphitised carbon grains are decimated byshocks or other energetic processes).

Benzene has been detected in the upper atmospheres of Jupiter, Saturn andTitan (e.g., Bezard et al. 2001; Waite et al. 2007, 2005), environments whereion-molecule reactions dominate over the neutral-neutral reactions which are themainstay of combustion chemistry. Similarly, if benzene were to form in the inter-stellar medium (ISM) or the intensely-irradiated environments around post-AGBstars, we would expect it to form via these ion-molecule reactions. Indeed, early

P.M. Woods: The Formation of Benzene in Dense Environments 237

models of dense interstellar clouds by McEwan et al. (1999) suggested that benzenewas formed thus:

C4H+3 + C2H2 −→ c−C6H+

5 + hν

(C4H+3 + C2H3 −→ c−C6H+

5 + H)c−C6H+

5 + H2 −→ c−C6H+7 + hν

c−C6H+7 + e− −→ c−C6H6 + H.

This model produced a fractional abundance of benzene of ≈10−9, several ordersof magnitude below that detected in CRL 618. In order, then, to explain theobserved abundance and investigate the formation of benzene in CRL 618, we willadopt an interstellar-based chemistry (Le Teuff et al. 2000; Millar et al. 2000), anda suitable physical model.

3 The formation of benzene in pre-planetary nebulae

(a) From Herpin & Cernicharo (2000) (b) Fractional abundance of C6H6 & progenitors.

Fig. 1. Physical and chemical models of CRL 618.

Herpin & Cernicharo (2000) were able to determine various physical propertiesof the circumstellar regions around CRL 618, and some of these are shown inFigure 1a. We choose to model the slowly-expanding (5 km s−1) torus, since thehigh density of this region (n ∼1011 cm−3, initially) extinguishes the significantUV radiation (2×105 G0) from the central star and allows complex chemistry todevelop. We also enhance the cosmic-ray ionisation rate by a factor of 100 in orderto account for X-rays from the hot star. At this stage of evolution, mass-loss fromthe star has ceased, and so we must consider a time-dependent situation. In thethin-shell approximation, we obtain the results shown in Figure 1b.

Benzene in the model is approximately 40 times less abundant than acetylene(a ratio observed by Cernicharo et al. 2001a) until about 1 300yr. At this stage,


dilution of the expanding torus means that photodestruction of the molecular mat-ter dominates. The major route to the formation of benzene in this environmentis similar to that described in Section 2, except that the route to C4H+

3 is simpler,viz.,

HCO+ + C2H2 −→ C2H+3 + CO

C2H+3 + C2H2 −→ C4H+

3 + H2

C2H+2 + C2H2 −→ C4H+

3 + H

(cf., McEwan et al. 1999). This is due to large fractional abundances of HCO+

(which is generally not abundant in AGB star envelopes but is abundant in thepost-AGB phase because of the more intense UV flux from the star) and C2H2

(which is generally not abundant in the ISM). Combined with a high degree ofUV-shielding and reasonable residence times, large amounts of benzene can form.For further details, see Woods et al. (2003, 2002).

4 The formation of benzene in protoplanetary discs

The inner regions of the protoplanetary accretion discs around young stars havea similar temperature and density to the torus modelled in the previous section.In this case, the physical model we use is significantly more complex (see Woods& Willacy 2009, for a complete description) but the chemistry is very similar,although we use solar, oxygen-rich initial abundances, and incorporate gas-graininteractions and grain-surface reactions into the chemical modelling.

Results of the model show that benzene can form at a reasonably high frac-tional abundance (∼10−6 cm−3) in the inner 3 AU of a protoplanetary disc. This issurprising, since protoplanetary discs are not carbon-rich environments; however,the cool midplanes of these discs allow water vapour to freeze out onto the surfaceof dust grains, effectively removing a major repository of oxygen from the gas-phase chemistry. Small hydrocarbons can then dominate the chemistry until theadvecting material is sufficiently warmed to liberate its oxygenous icy mantles.Interestingly, benzene forms in this environment via an odd-carbon-atom path-way rather than through the even-carbon-atom pathway (i.e., acetylene addition)discussed previously:

C3H4 + C3H+4 −→ c−C6H+

7 + H (4.1)c−C6H+

7 + grain −→ g−C6H6 + g−H. (4.2)

and this ion-molecule pathway is more efficient than a suggested three-body neutral-neutral reaction:

C3H3 + C3H3 + M −→ c−C6H6 + M

The abundance of benzene in the gas phase depends on the binding energy ofbenzene to the grain surface on which it resides after neutralisation (reaction 4.2).

P.M. Woods: The Formation of Benzene in Dense Environments 239

Fig. 2. The fractional abundance of gas-phase benzene in protoplanetary discs, using a

low (left) and high (right) value of the binding energy. Dashed lines show the approximate

disc surface.

Estimates of this binding energy vary, and thus two extreme cases (4750K and7580K) are shown in Figure 2 for the distribution of gas-phase benzene. A recentexperimental result suggests a value of 5600K (Thrower et al. 2009) is appropriate,similar to the binding energy of water. The reaction between C3H4 and its ion(reaction 4.1) has been measured in the laboratory, and has been shown to occurfor the main isomers of C3H4. These results are discussed in more detail in Woods& Willacy (2007).

5 The formation of PAHs in dense environments

The formation of benzene is thought to be the rate-limiting step in the formationof PAHs, and there is evidence to support this (Wang & Frenklach 1994). Po-tentially, then, once benzene is formed, PAHs could rapidly form from it, or fromrelated species like phenyl. Second-ring closure (e.g., to form naphthalene fromethylbenzene or similar) is thought to involve an energy barrier which requires tem-peratures of 900–1100K to overcome (Frenklach & Feigelson 1989). Temperatureslike this are present in the upper regions of protoplanetary discs, and material fromthe midplane of these discs is mixed to the surface on a timescale of ∼625yr (at3AU; Ilgner & Nelson 2006). Vertical mixing is neglected in the model of Woods &Willacy (2007), but we can estimate a formation timescale of a generic PAH (saywith 40 carbon atoms – a size which is stable to photodissociation) from benzene.Two-carbon (acetylene) units can be added at a rate of one every three years,assuming a density of ∼1010 cm−3 and a conservative association rate. Adding 17of these units to form a stable PAH would take 50 yr, significantly shorter thanthe mixing timescale calculated by Ilgner & Nelson (2006). Thus it seems reason-able that PAHs could form in protoplanetary discs. Of course, a more detailedtreatment is needed, and work is in progress on this topic.

Models of hydrocarbon chemistry in dense environments such as those dis-cussed show that there is a disposition towards the formation of complex species,


including benzene and PAHs. It may be that the destruction mechanisms of thesecomplex molecules are not fully understood, but even if the carbon skeletons ofmolecules are shattered by some energetic process, the fragments themselves arethe constituent building-blocks of PAHs, and can be re-incorporated into the for-mation process (Nı Chuimın 2009).

PMW would like to thank Tom Millar and Karen Willacy, who have contributed to this workover the past few years.

References

Bernard-Salas, J., Peeters, E., Sloan, G.C., et al., 2006, ApJ, 652, L29

Bezard, B., Drossart, P., Encrenaz, T., & Feuchtgruber, H., 2001, Icarus, 154, 492

Cau, P., 2002, A&A, 392, 203

Cernicharo, J., Heras, A.M., Tielens, A.G.G.M., et al., 2001a, ApJ, 546, L123

Cernicharo, J., Heras, A.M., Pardo, J.R., et al., 2001b, ApJ, 546, L127


Frenklach, M., 2002, Phys. Chem. Chem. Phys., 4, 2028


Geers, V.C., Augereau, J.-C., Pontoppidan, K.M., et al., 2006, A&A, 459, 545

Herpin, F., & Cernicharo, J., 2000, ApJ, 530, L129

Ilgner, M., & Nelson, R.P., 2006, A&A, 445, 223

Le Teuff, Y.H., Millar, T.J., & Markwick, A.J., 2000, A&AS, 146, 157

McEwan, M.J., Scott, G.B.I., Adams, N.G., et al., 1999, ApJ, 513, 287

Millar, T.J., Herbst, E., & Bettens, R.P.A., 2000, MNRAS, 316, 195

Nı Chuimın, R., 2009, Ph.D. Thesis, Univ. of Manchester

Thrower, J.D., Collings, M.P., Rutten, F.J.M., & McCoustra, M.R.S., 2009, MNRAS,394, 1510

Waite, J.H., Young, D.T., Cravens, T.E., et al., 2007, Science, 316, 870

Waite, J.H., Niemann, H., Yelle, R.V., et al., 2005, Science, 308, 982

Wang, H., & Frenklach, M., 1994, J. Phys. Chem., 98, 11465

Woods, P.M., & Willacy, K., 2009, ApJ, 693, 1360

Woods, P.M., & Willacy, K., 2007, ApJ, 655, L49

Woods, P.M., Nyman, L.-A., Schoier, F.L., et al., 2005, A&A, 429, 977

Woods, P.M., Millar, T.J., Herbst, E., & Zijlstra, A.A., 2003, A&A, 402, 189

Woods, P.M., Millar, T.J., Zijlstra, A.A., & Herbst, E., 2002, ApJ, 574, L167


EXPERIMENTAL STUDIES OF THE DISSOCIATIVERECOMBINATION PROCESSES FOR THE C6D

+6 AND C6D

+7

IONS

M. Hamberg1, E. Vigren1, R.D. Thomas1, V. Zhaunerchyk2, M. Zhang3,S. Trippel4, M. Kaminska5, I. Kashperka1, M. af Ugglas1, A. Kallberg6,A. Simonsson6, A. Paal6, J. Semaniak5, M. Larsson1 and W.D. Geppert1

Abstract. We have investigated the dissociative recombination (DR)of the C6D

+6 and C6D

+7 ions using the CRYRING heavy-ion storage

ring at Stockholm University, Sweden. The dissociative recombinationbranching ratios were determined at minimal collision energy, show-ing that the DR of both ions was dominated by pathways keeping thecarbon atoms together in one product. The absolute reaction crosssections for the titular ions are best fitted by σ(Ecm[eV]) = 1.3 ± 0.4×10−15 (Ecm[eV])−1.19±0.02 cm2 (C6D

+6 ) and σ(Ecm[eV]) = 1.1 ± 0.3×

10−15(Ecm[eV])−1.33±0.02 cm2 (C6D+7 ) in the intervals 3-300 meV and

3-200 meV respectively. The thermal rate constants of the titular reac-tions are best described by: k(T) = 1.3 ± 0.4 × 10−6(T/300)−0.69±0.02

cm3s−1 for C6D+6 and k(T)=2.0± 0.6 ×10−6 (T/300)−0.83±0.02 cm3s−1

for C6D+7 . These expressions correlates well with earlier flowing after-

glow studies on the same process.

1 Introduction

Polyaromatic hydrocarbons (PAHs) are thought to play a central role in astrochem-istry. It is suggested that PAHs may contain up to 20% of the available carbon

1 Department of Physics, Stockholm University, SE-106 91 Stockholm, Sweden;e-mail: [email protected] Radboud University Nijmegen, Institute for Molecules and Materials, PO Box 9010, 6500GL Nijmegen, The Netherlands3 Institute of Modern Physics, Chinese Academy of Sciences, 509 Nanchang Rd., Lanzhou730000, China4 Department of Physics, Albert-Ludwigs Universitat Freiburg, Hermann-Herder-Str. 3,79104 Freiburg, Germany5 Jan Kochanowski University, Swietokrzyska 15, 25406 Kielce, Poland6 Manne Siegbahn Laboratory, Frescativagen 26, 114 18 Stockholm, Sweden



in the ISM (Boulanger 1999). Whereas there is strong evidence for the presenceof PAHs in various astronomical objects, only the simplest aromatic compound,benzene (C6H6) has been unambiguously detected by the Infrared Space Observa-tory in the protoplanetary nebula CRL 618 (Cernicharo et al. 2001). Furthermore,benzene is present in the atmospheres of gaseous giant planets and has been foundin Titan’s atmosphere by the Composite Infra-Red spectrometer (CIRS) on boardthe Cassini-Huygens mission (Vinatier et al. 2007).

Benzene can function as template in the build-up of higher hydrocarbons in-cluding PAHs in the interstellar medium. In a recent paper by Le Page et al. (2009)it is suggested that PAHs may play a key role as catalyst in forming molecularhydrogen in the interstellar medium. This involves chemical trapping of H atomson dehydrogenated PAHs, ionisation of the PAH, followed by addition of anotherH atom with a subsequent release of H2 through dissociative recombination (DR).This sequence constitutes a circular processes in which a hydrogen molecule isformed from two atoms and the dehydrogenated PAH is recycled. Le Page et al.(2009) assume in their model branching fractions of 50% each for the two reactionchannels leading to H and H2 respectively in the DR of C6H+

7 .Investigation of the synthesis for benzene itself in these environments is there-

fore a key issue. In interstellar clouds as well as Titan’s atmosphere dissociativerecombination of the C6H+

7 ion (which is formed by a sequence iof ion-neutralreactions) leading to the unprotonated benzene has been invoked as the final stepin the synthesis of this compound (Woods et al. 2002; Vuitton et al. 2008):

C6H+7 + e− → C6H6 + H (1.1)

The DR process itself is, however, highly exoergic (ΔH = −5.73 eV for the above-mentioned reaction) and other reaction channels including ejection of carbon-containing fragments may be feasible. In addition, it has recently been shownthat ejection of a single H atom can be an only minor reaction channel for proto-nated molecules (Geppert et al. 2005; Hamberg 2010). Therefore knowledge of thebranching ratio of the DR of protonated benzene is crucial to validate the postu-lated formation mechanism in the interstellar medium and planetary atmospheres.This Paper presents an experimental study into the DR rate constant and productbranching fractions of C6D+

6 and C6D+7 .

2 Experiment

The experiments were carried out at the heavy-ion storage ring CRYRING atStockholm University, Sweden. The procedure has been explained in detail pre-viously (Neau et al. 2000) and will therefore only be summarised here. The ionswere formed by a discharge in a mixture of fully deuterated benzene and deuteriumgas in a hollow cathode type ion source. The gas was pulsed into the ion sourcereaching a peak pressure of ∼1 Torr in order to strive for collisional quenching.The ions were then accelerated by 40 kV and mass selected by the first bending

M. Hamberg et al.: DR of C6D+6 and C6D+

7 Ions 243

magnet in the injection line. After a 10-turn multi injection procedure into thestorage ring the ions were further accelerated by a RF-system to the maximumkinetic energy of ∼96/mion MeV (where mion is the ion mass in amu). A longstorage time of several seconds was ensured due to the low ring pressure of residualgas (typically ∼10−11 Torr). Thus the ions could be left to de-excite for 3 secondsbefore the measurement sequence started. This left ample time for vibrationallyexcited inner degrees of freedom to relax. In an electron cooler a cold, monoener-getic electron beam was overlapping the path of the ions over ∼85 cm which servedas DR interaction region. Neutral particles formed in the process were separatedfrom the ion beam, since they are unaffected by the dipole magnet located afterthe electron cooler and were recorded by an energy sensitive ion implanted silicondetector (IISD).

3 Data analysis and results

3.1 Branching fractions

Branching fraction measurements were performed at 0 eV nominal relative ki-netic energy. Neutral fragments deriving from DR-events impinge on the IISDat a timescale (ns) much shorter than the shaping time of the preamplifier (μs).Therefore the detected signal for each event will essentially reflect the sum of themasses of all the incoming fragments and thus yield no information about thebranching ratio. To come around this problem a transparent metal grid with atransmission probability of 0.297±0.015 is inserted in front of the IISD. With thegrid inserted all fragment masses appear in the energy spectrum and branchingratios can be derived from the intensity distribution of the different mass sig-nals. Fragment energy spectra with and without grid for C6D+

6 ion are shownin Figure 1. One can see a small signal around energy channel 950 indicatingthat some fragments containing 3 C-atoms are breaking loose from the rest ofthe molecule although the majority of DR processes seem to keep all six carbonatoms together. If splitting up of the ions into 2 fragments containing 3 carbonatoms would dominate the reaction flux the spectra measured with the grid in-serted would be severely different from the ones recorded without grid. The peakat full energy would then be smaller than the 3C peak found around channel 950due to the comparatively low transmissibility of the grid. No obvious peaks arefound at the energies corresponding to fragnments with two or four carbon atoms.The lack of smaller fragments is even more pronounced for the C6D+

7 ion wherethe spectrum structure essentially indicates that all fragments keep the six carbonatoms together and no other peaks are distinguishable (see Fig. 1). Rescaling theMCA-channels into mass units (which can only be done in a very approximativeway due to the broadness of the features) shows that the difference of the peakcenters for the peaks corresponding to the highest masses in Figure 1 with thegrid in and out is in the order of one amu. This is well in line with ejection of oneor two deuterium atoms and a preservation of the aromatic ring structure during


500 1000 1500 2000

C6D

7

+

Inte

nsity

/

a.u

.

Fragment energy / MCA-channel

C6D

6

+

Inte

nsity

/

a.u

.

Fig. 1. MCA spectra from branching fraction measurements of the C6D+6 (top) and

C6D+7 (bottom) ions. The data obtained without grid is depicted with a dotted line

whereas the ones recorded with the grid inserted are represented with a solid line.

DR. However, further analysis were complicated due to losses of fast deuteriumfragments bypassing the active detector area. Possible losses of heavier fragmentscontaining carbon atoms were estimated to be well below 10% using Monte Carlosimulations in a similar fashion as Thomas (2008) and should therefore not affectthe obtained spectra much.

3.2 Cross section and rate constant

To measure the DR rate constant the electron velocity was ramped relative to theion velocity by scanning the electron acceleration voltage in the electron cooler.The neutral particles deriving from the DR-processes were detected by the IISDand counted by a multichannel scaler (MCS) using 2 ms time bins. Each bincorresponds to an interaction energy in the interval between ∼2–1000 meV. Back-ground processes from ions interacting with residual gas in the storage ring weresimultaneously collected by a micro channel plate (MCP) located at the end of astraight section and recorded by a MCS. The background signal intensity is di-rectly proportional to the ion current and could therefore be scaled to the absoluteion current which was measured with a capacitive pick-up calibrated by an ACtransformer directly after acceleration (Paal et al. 2006). Therefore the mean ioncurrent throughout the ring-cycle could be determined and fitted by a decay curve.This fit was also used for background subtraction of the DR MCS spectrum before


7 Ions 245

applying the following formula for calculation of the rate coefficient:

k = (dN

dt)vivee

2r2eπ

IeIil, (3.1)

where dN/dt is the measured DR-count rate, Ie, ve and Ii, vi are the electronand ion currents and velocities (in the laboratory frame), respectively and e is theelementary charge, re is the radius of the electron beam and l is the interactionregion length.

Corrections to the data had to be made for: (i) Space charge effects wherethe electrons experience a drop in the acceleration potential at the cathode in thecooler due to free electrons traveling ahead of them. (ii) Toroidal effects derivingfrom ion electron collisions at higher energies than the detuning energy in theregions where the electron beam is bent into and out from the interaction region.(iii) The electrons have a non negligible transversal energy spread of 2 meV. Therate constant can therefore be expressed as below through which the DR crosssection can be deconvoluted using Fourier methods (Mowat et al. 1995):

k =∫ ∞

0

vrelf(v⊥)σ(vrel) dv⊥, (3.2)

where f(v⊥) is the transversal electron velocity distribution in the center of massframe, v⊥ the transversal electron velocity and vrel the relative velocity respec-tively. The result for the C6D6+ ion is displayed in Figure 2 together with the bestfit between 3–300 meV giving a cross section of σ(Ecm[eV]) = 1.3 ± 0.4 × 10−15

(Ecm[eV])−1.19±0.02 cm2. Applying the same procedure for the C6D+7 ion (see

Fig. 2) resulted in a best fit over the interval 3–200 meV of: σ(Ecm[eV]) = 1.1±0.3 × 10−15(Ecm[eV])−1.33±0.02 cm2. The thermal rate constant can be obtainedfrom the cross sections by applying the formula:

k(T ) =8πme

(2πmekBT )3/2

∫ ∞

0

Ecmσ(Ecm)e−Ecm/kBT dEcm, (3.3)

where me is the electron mass, kB is Boltzmann’s constant, T is the electron tem-perature and Ecm is the centre of mass energy. The obtained thermal rates are:k(T) = 1.3 ± 0.4 × 10−6(T/300)−0.69±0.02 cm3s−1 for C6D+

6 and k(T) = 2.0 ±0.6 × 10−6 (T/300)−0.83±0.02 cm3s−1 for C6D+

7 . The systematical errors are arisingfrom uncertainties in the ion and electron currents and the length of the interactionregion electron density. These errors combined with the statistical uncertaintiesare estimated to be 25% and 30% for C6D+

6 and C6D+7 respectively.

4 Discussion

For both ions DR experiments have previously been performed using flowing af-terglow (FA) technique. The thermal rates at 300 K was found to be 1.0 ± 0.3 ×10−6 cm3 s−1 (C6D+

6 ) and 2.4± 0.4× 10−6 cm3 s−1 (C6D+7 ), which is in very good


1E-5 1E-4 1E-3 0.01 0.1 1

1E-16

1E-15

1E-14

1E-13

1E-12

1E-11

1E-10

1E-9

1E-8

C6D

7

+

Cro

ss S

ectio

n /

cm

2

Center Of Mass Energy / eV

1E-5 1E-4 1E-3 0.01 0.1 1

1E-16

1E-15

1E-14

1E-13

1E-12

1E-11

1E-10

1E-9

1E-8

C6D

6

+

Cro

ss S

ectio

n /

cm

2

Fig. 2. Cross section of the DR of C6D+6 (top) and C6D

+7 (bottom) versus relative kinetic

energy. The triangles shows individually measured spots while the dashed line shows the

best fit between 3–300 meV (C6D+6 ) and 3–200 meV (C6D

+7 ) respectively.

agreement with previous results for the undeuterated isotopomers (Abouelazizet al. 1993; McLain et al. 2004). The lowest formation enthalpy for the undeuter-ated C6H+

6 and C6H+7 ions are found for their cyclic forms and are 975.8 kJ mol−1

and 854 kJmol−1 respectively (Lias et al. 1988). Abouelaziz et al. (1993) found intheir DR experiment of C6H+

6 no difference in the results when they formed theion through charge transfer reaction with C+, Xe+ and Kr+ suggesting that thestructures of the product ions were the same in each case. These charge transferprocesses are exoergic with 2 eV, 2.88 eV and 4.75 eV respectively, which canbe compared to at least 2.8 eV needed for breaking the benzene ring. Even athigher exoergicities ring opening seems therefore not to occur. In the ion sourcethe C6D+

6 ions could be formed through charge transfer from D+2 and D+ (exoer-

gic with 6.17 eV and 4.34 eV, respectively (Lias et al. 1988)) or charge transferreaction with D+

3 leading to D2 and D (close to thermoneutral, i.e. with a reac-tion enthalpy close to zero). Electron impact and Penning ionization are otherprocesses contributing and one can not rule out formation of other isomers. Nev-ertheless, we assume that the cyclic structure should dominate the stored ions. Inall these thermodynamic considerations the enthalpies of the undeuterated com-pounds were considered due to the lack of data for the heavier isotopomeres.

In our set-up the C6D+7 ions should predominantly be formed through protona-

tion by D+3 and to some extent reactions with D+

2 and D+ (exoergic with 3.48 eV,


7 Ions 247

5.17 eV and 7.87 eV respectively, Lias & Ausloos 1985). It is also found that theion can be formed by hydrogen atom abstraction from hydrogen donors. Flowingafterglow experiments were carried out by Milligan et al. (2002) and Spanel et al.(1995) in which they found that reaction of H+

3 with benzene leads to C6H+7 with

unity probability which was described as “a manifestation of the extraodrinarystability of this protonated molecule”. Unfortunately however, their type of setupwas not sensitive to structural isomers (Spanel & Smith 1998). This is somewhatdisadvantageous, since also stable five membered ring and open-chain isomers ofC6D+

7 exist. Kuck (1990) claimed that the cyclic C6H+7 was, beyond any doubt,

the most stable light hydrogen isomer. In a ICR study, even reactions betweenaliphatic (neutrals e.g. C3H4) and ions (C3H+

4 ) resulting in the formation of C6H+7

at least partly lead to protonated cyclobenzene (Lias & Ausloos 1985). This showsthat even if the formation process of C6D+

7 in our ion source to some extent resultsin open-chain C6D+

7 isomers, ring closure of these species to deuteronated benzenecould be possible. Although we cannot exclude different isomeric forms we findit likely that the aromatic form should be dominating. In addition, it shouldbe noted that even protonated benzene has two isomers – a σ and a π-complex.However, Solca & Dopfer (2002) found no presence of the π-complex in a spectro-scopic study of C6H+

7 ions produced by proton transfer to benzene from H+3 . This

is probably also the major formation process of the C6H+7 ion in our experiment,

thus we assume the ion present in our experiment to be the σ-C6H+7 .

Thus, our findings are in line with a formation of benzene in the interstellarmedium, circumstellar envelopes and planetary ionospheres through ion-neutralreactions leading to C6H+

7 followed by DR yielding benzene. Also the rate constantpreviously measured in flowing afterglow experiments (Abouelaziz et al. 1993;McLain et al. 2004) and used in models of Titan’s ionosphere (Vuitton et al. 2008)has been corroborated in our study, albeit the temperature dependence of the DRrate of C6D+

7 has been found to be considerably weaker (the exponential factorbeing −0.83 instead of −1.3). This could be due to a different reactivity of groundand rovibrationally excited states of the ion. In our storage ring, as opposedto afterglow experiments, the distribution between these states is not changedthroughout the energy scan. Such different exponential factors of temperaturedependence lead to considerably unlike rates at very low temperatures and it hasto be seen how model predictions of benzene abundances in very cold environmentslike dark clouds are affected by that. The effect of the new rate constants onmodels of warmer environments like planetary ionospheres are, however, probablynegligible.

Our results on the DR of C6D+7 pose a challenge to neither the invoked forma-

tion mechanism of benzene nor to the proposed H2 formation catalyzed by PAHs.They can, on the contrary, not serve to corroborate them, since we could not estab-lish the branching ratio of the ejection of H2 and the formation of benzene. If thisreaction would only result in production of H2, it would render the invoked reac-tion sequence of benzene formation unfeasible. If on the other hand the branchingratio of the formation of benzene would be close to 1, the catalytic cycle of H2

formation postulated by Le Page et al. (2009) would be impossible. Experiments


on storage rings with higher rigidity, which would enable us to distinguish betweenH2 and H ejection are therefore indespensable to clarify the validity of those tworeaction sequences.

5 Conclusion

Storage ring technique has for the first time been employed for DR measurementsof the simplest aromatic hydrocarbon ions C6D+

6 and C6D+7 . The rate constants

are in line with previous flowing afterglow measurements at 300 K. Although itwas not possible to fully elucidate the branching fractions between the reactionchannels, our studies showed that the C6-structure is retained almost entirely inboth cases. Dissociative recombination of C6H+

7 therefore still can be regardedas a feasible final step in the synthesis of benzene the interstellar medium andplanetary ionospheres. Our findings also do not disprove the molecular hydrogenformation mechanism by PAHs proposed by Le Page et al. (2009).

We thank the staff at the Manne Siegbahn Laboratory for making these experiments possibleby excellent technical support. W.D.G. thanks the Swedish Research Council for his Senior Re-searcher grant (contract number 2006-427) and the Swedish Space Board (grant number 76/06).M.K. thanks the Swedish Institute for financial support and also acknowledges support fromthe Ministry of Science and Higher Education, Poland, under contract N202 111 31/1194. S.T.acknowledges financing from the COST ACTION (CM0805).

References

Abouelaziz, H., Gomet, J.C., Pasquerault, D., Rowe, B.R., & Mitchell, J.B.A., 1993,J. Chem. Phys., 99, 237

Boulanger, F., 1999, in Solid State Matter: The ISO Revolution, Lecture 2, ed.L. dHendecour, C. Joblin, & A. Jones (New York Springer), 19

Cernicharo, J., Heras, A.M., Tielens, A.G.G.M., et al., 2001, Astrophys. J. Lett., 546,123

Geppert, W.D., Thomas, R.D., Ehlerding, A., et al., 2005, J. Phys. Conf. Ser., 4, 26

Hamberg, M., 2010, A&A, 514, A83

Kuck, D., 1990, Mass Spectrometry Reviews, 9, 187

Le Page, V., Snow, T.P., & Bierbaum, V.M., 2009, ApJ, 704, 274

Lias, S.G., & Ausloos, P., 1985, J. Chem. Phys., 82, 3613

Lias, S.G., Bartmess, J.E., Liebman, J.F., et al., 1988, J. Phys. Chem. Ref. Data., 17Suppl. 1

McLain, J.L., Poterya, V., Molek, C.D., Babcock, L.M., & Adams, N.G., 2004, J. Phys.Chem., 108, 6704

Milligan, D.B., Wilson, P.F., Freeman, C.G., Meot-Ner, M., & McEwan, M.J., 2002,J. Phys. Chem. A, 106, 9745

Mowat, J.R., Danared, H., Sundtroem, G., et al., 1995, Phys. Rev. Lett., 74, 50


7 Ions 249

Neau, A., Al-Khalili, A., Rosen, S., et al., 2000, J. Chem. Phys., 113, 1762

Paal, A., Simonsson, A., Dietrich, J., & Mohos, I., 2006, Proceedings of EPAC2006, 1196

Solca, N., & Dopfer, O., 2002, Angew. Chem. Int. Ed., 41, 3628

Spanel, P., & Smith, D., 1998, Int. J. Mass Spectrom., 181, 1

Spanel, P., Smith, D., & Henchman, M., 1995, Int. J. Mass Spectrom. Ion Proc., 141,117

Thomas, R.D., 2008, Mass Spectrom. Rev., 27, 485

Vinatier, S., Bezard, B., Fouchet, T., et al., 2007, Icarus, 188, 120

Vuitton, V., Yelle, R., & Cui, J., 2008, J. Geophys. Res., 113, E05007

Woods, P.M., Millar, T.J., Zijlstra, A.A., & Herbst, E., 2002, Astrophys J. Lett., 574,167


VUV PHOTOCHEMISTRY OF PAHS TRAPPED ININTERSTELLAR WATER ICE

J. Bouwman1, H.M. Cuppen1, L.J. Allamandola2 and H. Linnartz1

Abstract. The mid-infrared emission of Polycyclic AromaticHydrocarbons is found in many phases of the interstellar medium.Towards cold dense clouds, however, the emission is heavily quenched.In these regions molecules are found to efficiently freeze-out on inter-stellar grains forming thin layers of ices. PAHs are highly non-volatilemolecules and are also expected to freeze-out. PAHs trapped in in-terstellar ices are likely to participate in the overall chemistry, leadingto the formation of cations and complex molecules in the solid-state.The work presented here aims to experimentally study the chemical re-actions that PAHs undergo upon vacuum ultraviolet irradiation whentrapped in interstellar H2O ice.

1 Introduction

PAHs are abundantly present in many phases of the ISM as evidenced by their mid-IR emission bands. Towards dense clouds, however, this emission process is heavilyquenched. A possible explanation is that PAHs freeze-out on cold interstellargrains, much like small molecules such as H2O CO, CO2, and CH3OH, whichhave been identified in interstellar ices. Once trapped in ices, PAHs are expectedto absorb mid-IR photons, rather than emitting them.

Most PAH species have been studied spectroscopically in matrix isolation ex-periments. Pioneering work has been performed in the past to understand thechemical behavior of PAHs trapped in interstellar ice analogues; the formation ofnew PAH based products was studied by Bernstein et al. (1999) and the efficientionization of PAHs in H2O ice was first noted by Gudipati & Allamandola (2003).Here, we build on these first findings by employing near-UV/VIS absorption spec-troscopy to present ionization and photoproduct formation rate constants for a

1 Raymond and Beverly Sackler Laboratory for Astrophysics, Leiden Observatory, LeidenUniversity, PO Box 9513, 2300 RA Leiden, The Netherlands2 NASA-Ames Research Center, Space Science Division, Mail Stop 245-6, Moffett Field,CA 94035, USA



Fig. 1. A typical near-UV/VIS absorption spectrum of the PAH pyrene (Py) trapped in

H2O ice photolyzed for 900 s at a temperature of 125 K. Assignments of the destroyed

(negative signal) neutral Py bands and newly formed (positive signal) Py photoproduct

bands, the Py cation (Py+) and 1-hydro-1-pyrenyl radical (PyH) are indicated by arrows.

The inset shows a blow-up of the 350 to 490 nm range. Figure is taken from Bouwman

et al. (2010).

number of PAHs trapped in H2O ice. Additionally, the photoproducts are alsoidentified in-situ by mid-IR spectroscopy. The results are used to derive an upperlimit for the possible contribution of PAHs to the 5–8 μm absorption complextowards protostellar objects.

2 Experimental

The PAH:H2O ice photolysis experiments are performed on two different setups,which have been described in much detail in Bouwman et al. (2009) and Hudginset al. (1994) and will only be summarized here briefly. Both setups are basedon the same principle and only differ in sample preparation and spectroscopictechnique. The systems consist of a high-vacuum chamber pumped down to apressure of ∼10−7 mbar. A cold finger containing a sample window is suspendedin the vacuum chamber and is cooled down to a temperature of 15 K by meansof a closed cycle He refrigerator. The sample temperature can be controlled bymeans of resistive heating.

PAH:H2O ice samples are grown on the cold sample window by vapor co-depositing milli-Q grade H2O and the PAH under investigation. The PAHs an-thracene (Ant, C14H10, Aldrich, 99%), pyrene (Py, C16H10, Aldrich, 99%), orbenzo[ghi]perylene (BghiP, C22H12, Aldrich, 98%) for the mid-IR and additionallycoronene (Cor, C24H12, Aldrich, 99%) for the near-UV/VIS are used as commer-cially available and brought in the gas phase by heating a sample container toa temperature which is just high enough to create the desired PAH:H2O sample

J. Bouwman et al.: PAH:H2O Ice Photochemistry 253

Fig. 2. Time evolution of the relative column density of neutral pyrene (Py) and the

pyrene cation (Py+) in the ice as a function of photolysis time (VUV fluence). Based on

data from Bouwman et al. (2010).

ratio. Mixing ratios vary from 1:10 to 1:200 for the mid-IR experiments to 1:500to 1:10 000 for the optical studies.

The ices for both experimental methods are subsequently irradiated with avacuum ultraviolet (VUV) H2 microwave discharge light source to simulate theinterstellar radiation field. For the mid-IR spectra, ices are irradiated for 5, 10,15, 30, 50, 120, and 180 minutes and after each VUV dose one spectrum rangingfrom 4000 to 400 cm−1 is taken at 0.5 cm−1 resolution. In the near-UV/VISexperiment, spectra are recorded over the spectral range from 280 to 800 nm witha resolution of 0.9 nm.

3 Near-UV/VIS absorption spectroscopy

Near-UV/VIS absorption spectra are obtained for a set of VUV processedPAH:H2O ices (Bouwman et al. 2011a). A typical example of a spectrum isshown in Figure 1. This spectrum is obtained by taking the ratio of a spectrumafter photolysis time t and a spectrum taken of the freshly deposited ice ln(It/I0).The resulting spectrum exhibits positive absorptions arising from the newly formedspecies and negative absorptions arising from the consumed neutral parent PAHspecies. Clearly, new bands appear in the spectrum which can be mainly assignedto electronic transitions of the cationic PAH species.

Acquiring a spectrum, such as that displayed in Figure 1, takes 10 s and asmany as 1440 spectra are obtained over the course of a single 4 hour photolysis ex-periment. Currently, optical spectra are available for VUV irradiated Anthracene:H2O, Pyrene:H2O, BenzoghiPerylene:H2O, and Coronene:H2O (Bouwman et al.2009, 2010, 2011a) ice mixtures. These data allow us to derive the time evolutionof the column density of PAH and PAH+ species in the ice as indicated in Figure 2for a Py containing H2O ice. Both neutral parent PAH and PAH+ cation species


Fig. 3. Comparison between the mid-IR spectra of pyrene (Py) and pyrene cation (Py+)

in an Argon matrix (a and c) and the same spectra for Py and Py photoproducts trapped

in H2O ice (b and d). The top spectrum (e) indicates the subtraction spectrum (d−n×b)

which exhibits photoproducts that appear upon photolysis of a Py:H2O ice only. Figure

is based on data from Bouwman et al. (2011b).

are consumed at the end of an experiment. Together with the loss of the PAH andPAH+ species a broad absorption superposed on the baseline arises, which mostlikely makes up the absorption by the newly formed photoproducts.

A time-dependent chemical analysis is performed for all PAH species mentionedabove. A chemical model is fitted to the experimental data and the rate constantsfor processing PAHs in an interstellar ice analogue are derived. It is found that therate constant for ionization of PAHs in H2O ice is quite similar for all PAHs underinvestigation here. The derived average PAH ionization rate constant (kion ≈9×10−18 cm2 photon−1) can be used to quantify the importance of PAH ionizationin the interstellar case by implementing this rate constant in astrochemical models.

4 Mid-IR absorption spectroscopy

Mid-IR spectroscopy is performed on a set of VUV photolyzed PAH containingH2O ices (Bouwman et al. 2011b). Spectra are taken on the freshly deposited ice(spectrum X) and after a certain dose of VUV light (spectrum Y ). A typical spec-trum of a neutral PAH in H2O ice, photolyzed PAH in H2O ice and a photolyzedPAH in H2O ice minus the contribution of the remaining neutral PAH species, n(photoproducts solely = Y − nX) is displayed in Figure 3 for Py:H2O ice. Thefactor (1 − n) directly reflects the amount of consumed neutral species.

The spectra of the photoproducts that appear upon photolysis of the PAHin H2O ice are compared to known spectra of PAH+ species measured in an Ar

J. Bouwman et al.: PAH:H2O Ice Photochemistry 255

Fig. 4. Comparison between the residuals in the 5–8 μm spectral region towards W33A

(Keane et al. 2001) and the photoproducts formed by long duration photolysis of a

pyrene containing H2O ice. The dotted line guides the eye to the 6.2 μm interstellar

unidentified residual absorption and the cluster of the main pyrene:H2O photoproduct

bands.

matrix (Hudgins et al. 1995a, 1995b). In this way, some of the absorption bandscan be directly assigned to mid-IR absorptions of the PAH cation. Other bands,however, must be caused by different band carriers.

A comparison of the new absorption bands originating from unknown specieswith typical mid-IR band positions indicates that the species responsible for theabsorptions are likely the parent PAH with side-groups attached. In agreementwith an off-site analysis of the photoproduct species resulting from the photolysisof PAH:H2O ices (Ashbourn et al. 2007; Bernstein et al. 1999), we attribute theseabsorptions to hydrogenated, hydroxylated and oxydized PAH species.

5 Astrophysical implications

Many interstellar ices have been unambiguously identified by their mid-IR absorp-tion bands observed towards protostellar objects or background stars behind denseclouds. Most of these species are rather simple molecules, such as H2O, CO, CO2,NH3, and CH3OH (Gibb et al. 2000). PAHs frozen out on interstellar grains havenot yet been unambiguously identified. Here we derive from the presented spectraan upper limit for the contribution that PAH species may have to the complexabsorption in the 5–8 μm mid-IR spectral range towards both high- and low-massprotostars.


The 5–8 μm spectral region is dominated by the strong H2O bending mode.The residuals that arise after subtraction of the contribution of the H2O bendingmode show that more species are responsible for the overall absorption. As canbe seen in Figure 4, there is a clear overlap between the residuals of W33A (takenfrom Keane et al. 2001) and the photoproduct bands that appear upon prolongedirradiation of a PAH:H2O ice mixture. A band strength can be estimated forthe PAH:H2O photoproduct, because the number of molecules responsible for thisband is known. This band strength is used to determine the number of moleculesbeing responsible for the 6.2 μm band (Fig. 4) if these absorption bands are indeedcaused by PAH:H2O photoproducts. In Bouwman et al. (2011b), an upper limitof 2–3% with respect to H2O is derived for the contribution of PAHs to the 5–8 μmabsorption complex.

This work is financially supported by NASA’s Laboratory Astrophysics and AstrobiologyPrograms, “Stichting voor Fundamenteel Onderzoek der Materie” (FOM), and “the NetherlandsResearch School for Astronomy” (NOVA). J. Bouwman gratefully acknowledges the “Search forExtraterrestrial Intelligence” (SETI) institute for financial support.

References

Ashbourn, S.F.M., Elsila, J.E., Dworkin, J.P., et al., 2007, Meteorit. Planet. Sci., 42,2035

Bernstein, M.P., Sandford, S.A., Allamandola, L.J., et al., 1999, Science, 283, 1135

Bouwman, J., Cuppen, H.M., Allamandola, L.J., & Linnartz, H., 2011a, A&A, accepted

Bouwman, J., Cuppen, H.M., Bakker, A., Allamandola, L.J., & Linnartz, H., 2010, A&A,511, A33+

Bouwman, J., Mattioda, A.L., Allamandola, L.J., & Linnartz, H., 2011b, A&A, 525, A93

Bouwman, J., Paardekooper, D.M., Cuppen, H.M., Linnartz, H., & Allamandola, L.J.,2009, ApJ, 700, 56

Gibb, E.L., Whittet, D.C.B., Schutte, W.A., et al., 2000, ApJ, 536, 347

Gudipati, M.S., & Allamandola, L.J., 2003, ApJ, 596, L195

Hudgins, D.M., & Allamandola, L.J., 1995a, J. Phys. Chem., 99, 3033

Hudgins, D.M., & Allamandola, L.J., 1995b, J. Phys. Chem., 99, 8978

Hudgins, D.M., Sandford, S.A., & Allamandola, L.J., 1994, J. Phys. Chem., 98, 4243

Keane, J.V., Tielens, A.G.G.M., Boogert, A.C.A., et al., 2001, A&A, 376, 254

PAHs in Regions

of Planet Formation


OBSERVATIONS OF HYDROCARBON EMISSIONIN DISKS AROUND YOUNG STARS

B. Acke1, 2

Abstract. I give an overview of the recent scientific results based onobservations of PAH emission from circumstellar disks around youngstars. The stellar radiation field plays a key role in the excitation anddestruction of the PAH molecules in the disk. The detection rate ofPAH emission in disks is optimal for stars of spectral type A. Aroundstars of similar temperature, the disk structure determines the PAHemission strength: disks with a flared geometry produce stronger PAHemission than flattened disks. The spectral properties of the emissionfeatures, indicative of the chemistry of the emitting hydrocarbons, isclosely linked to the central star radiation field. The main PAH featuresshift to redder wavelengths with decreasing stellar effective tempera-ture. This trend has been interpreted as an indication for a higheraliphatic/aromatic ratio of the hydrocarbon mixture around cool starswith respect to hot stars. An alternative explanation may be a moresignificant contribution to the infrared emission of very small grainsaround cooler stars.

1 Introduction

As a result of the star-formation process, young stars are surrounded by circum-stellar disks. These disks, rich in gas and dust, are believed to be the birthplacesof planetary systems. The disk dissipation timescale is short; In only a few millionyears, half of the newly formed stars become diskless (e.g., Fedele et al. 2010). Itis in this short time span that processes such as dust grain growth, crystallisation,sedimentation, photoevaporation, and chemical evolution can take place beforethe dust is removed. They affect the appearance of the system throughout theelectromagnetic spectrum.

1 Instituut voor Sterrenkunde, University of Leuven, Celestijnenlaan 200D, 3001 Leuven,Belgium; e-mail: [email protected] Postdoctoral Fellow of the Fund for Scientific Research, Flanders



Young stars are traditionally categorized according to their spectral type.T Tauri stars are young analogues of the Sun, with spectral type later thanmid-F and main-sequence masses M� ≤ 1.5 M�. Higher-mass young stars arecalled Herbig Ae/Be stars (Herbig 1960). Observations indicate that T Tauri andHerbig Ae stars follow a similar disk evolution. Evidence is growing, however, thatdisks around Herbig Be stars (M ≥ 3 M�, spectral type earlier than mid-B) arequalitively different in structure. In the following, it will become clear that theT-Tauri, Herbig-Ae and Herbig-Be classification is also meaningful in the frame-work of the hydrocarbon emission.

In this Review, I group the scientific results on PAHs in young circumstellardisks in three sections. Section 2 focuses on the detection rates of disk PAHemission features for different types of young stars. In Section 3, it is shownthat the strength of the PAH emission is linked to the disk geometry. Finally,I summarize the chemical properties of the hydrocarbons, in relation to the stellarradiation field, in Section 4.

2 PAH detection rates in disks

The detection rate of IR emission features produced by PAH molecules in thecircumstellar disk is low for T Tauri stars (∼10%) and Herbig Be stars (<35%).More than two thirds of the intermediate-mass Herbig Ae stars, on the other hand,have detected PAH emission from disks.

2.1 T Tauri stars

Geers et al. (2006) and Geers (2007) searched for PAH emission in the spectraof 37 T Tauri stars obtained with the Infrared Spectrograph aboard the SpitzerSpace Telescope (Spitzer-IRS; Houck et al. 2004). They determined a detectionrate of 11–14%, with no detections for stars with spectral type later than G8.Furlan et al. (2006) studied Spitzer-IRS spectra of young T Tauri disks in theTaurus star-forming region, whereas Oliveira et al. (2010) investigated youngstellar objects in the Serpens Molecular Cloud. Only 3–4% of the disk sourceshave detected PAH features. One could think that the weak UV field of a T Tauristar is insufficient to excite the PAH molecules in the disk. However, Geers et al.(2006) compare their data to a simple model for PAH emission in a circumstellardisk. They assume PAH molecules containing 100 carbon atoms (NC=100). Toexplain the low detection rate, the authors conclude that the abundance of gas-phase PAHs in disks around T Tauri stars must be 10–100 times lower than thePAH abundance of the general interstellar medium (ISM; Cesarsky et al. 2000;Habart et al. 2001).

Geers et al. (2009) searched for PAH emission in low-mass embedded proto-stars, the evolutionary stage preceding the T Tauri stage, and determined an evenlower detection rate (2%). The authors argue that absorption by the envelopesurrounding the star and disk is not sufficient to explain the absence of features,

B. Acke: PAHs in Circumstellar Disks 261

and again suggest that lower PAH abundances are needed. They hypothesizethat PAHs enter the disk either frozen out on dust grains, or in agglomerates.Quanz et al. (2007) studied the Spitzer-IRS spectra of a sample containing 14 FUOrionis stars. The latter are young low-mass objects that undergo energetic accre-tion events, leading to strong outbursts at optical and near-IR wavelengths. Onlyone object, possibly even misclassified as a FU Orionis star, shows hydrocarbonemission features, amongst others the feature at 8.2μm. The low detection rate inthis sample (≤7%) supports the hypothesis of PAH underabundance in the disksaround young low-mass stars.

Siebenmorgen & Krugel (2010) modeled the effect of hard radiation on PAHmolecules in a circumstellar disk (see also Siebenmorgen et al., elsewhere in thisvolume). The authors find that for typical T Tauri X-ray luminosities (LX/L� ∼10−4–10−3; Preibisch et al. 2005), PAH molecules in the disk surface are efficientlydestroyed at all radial distances, leading to a suppression of the PAH emission.Siebenmorgen & Krugel (2010) suggest that in the few T Tauri disks with de-tected PAH emission, turbulent vertical mixing brings PAHs to the upper disklayers faster than photodestruction removes the molecules. The hard radiationcomponent in Herbig Ae stars is much weaker (LX/L� < 10−7 with a few excep-tions; Stelzer et al. 2006). A lower destruction rate is expected, consistent withthe much higher detection rates (see Sect. 2.3).

However, it may be that the low detection rate is biased because of the focus onthe 6.2 and 11.3μm features. Bouwman et al. (2008) find that 5 out of the 7 low-mass young stars in their sample display hydrocarbon emission features. Thefeatures are weak compared to the underlying continuum and silicate emission,and a dedicated fit is required to reveal them. Bouwman et al. show that thehydrocarbon emission spectrum is different from the spectrum seen in higher-massyoung stars. It is dominated by a feature at 8.2μm, and has weak 6.2 and 11.3μmfeatures. This deviating hydrocarbon spectrum in T Tauri stars is consistent withthe picture outlined in Section 4.

2.2 Herbig Be stars

The infrared spectra of roughly half of the Herbig Be stars contain PAH emissionfeatures (Acke and van den Ancker, unpublished). Due to the strong UV field of thecentral star, however, most of the PAH emission comes from surrounding clouds,remnants of the natal cloud, rather than from the disk. Verhoeff et al. (A&Asubmitted) have studied the extended PAH emission in a sample of 30 Herbig Bestars. None of the emission can unambiguously be attributed to the disk, butin 35% of the targets, PAH emission from the disk cannot be excluded either.Verhoeff et al. argue that the disks around Herbig Be stars are fundamentallydifferent from those around lower-mass stars in terms of disk geometry: they areradially less extended and vertically flatter. The authors state that, even if PAHmolecules are present in the disk surface at ISM abundances, their emission wouldbe insignificant due to the disk geometry, and swamped by the PAH emission from


the extended cloud. Moreover, the hard radiation field of early-type young starsmay shift the PAH destruction radius beyond the disk outer radius, suppressingthe PAHs in the disk altogether.

2.3 Herbig Ae stars

The intermediate-mass Herbig Ae stars have by far the highest detection rate ofPAH emission in disks. Their stellar radiation field is optimal to excite but not de-stroy the molecules. Acke & van den Ancker (2004) studied the infrared spectra of46 Herbig Ae/Be stars, mainly based on data obtained with the Short-WavelengthSpectrometer aboard the Infrared Space Observatory (ISO-SWS; de Graauw et al.1996; Kessler et al. 1996). The detection rate of PAH emission features is 57%.Recently, Acke et al. (2010) investigated a sample of 53 Herbig Ae stars, observedwith Spitzer-IRS. The authors detect PAH features in 70% of the targets. TheISO-SWS and Spitzer-IRS sample overlap in 27 sources. A closer look at this sub-sample shows that the higher detection rate in the Spitzer-IRS spectra is simplydue to the superior sensitivity of the latter instrument.

Because of the high detection rate in Herbig Ae stars, most studies of PAHs inyoung disks focus on these targets.

3 Emission strength and disk geometry

Spatially resolved observations of PAH emission around young stars have clearlydemonstrated the disk origin of the emission. Two techniques are commonly used:long-slit spectroscopy (e.g., van Boekel et al. 2004; Habart et al. 2004b, 2006;Geers et al. 2007) and narrow-band imaging (e.g., Lagage et al. 2006; Doucet et al.2007). The focus lies on the PAH features that are accessible from the ground, i.e.the 3.3 (L-band), 7.7, 8.6 and 11.3μm (N-band) features. The PAH emission regionhas a typical extent of tens of AU for the 3.3μm feature to hundreds of AU forthe features at longer wavelengths. Habart et al. (2004a) and Visser et al. (2007)have modeled PAH emission from a circumstellar disk. These authors concludethat large (i.e., NC=100 as opposed to NC=30) PAHs are needed to explain theobserved spatial extent.

Lagage et al. (2006) & Doucet et al. (2007) have imaged the HD97048 disk innarrow-band filters containing the 8.6 and 11.3μm PAH features, and at continuumwavelengths in between, using VLT/VISIR (Lagage et al. 2004). They confirm thefindings mentioned above. Moreover, the images clearly show the flared geometryof the disk surface: the disk opening angle increases with radial distance to thestar, which gives it a bowl-shaped appearance (see Fig. 1). The measured increaseof surface height above the midplane with radius follows H ∝ R1.26±0.05, in perfectagreement with the predicted value for models in hydrostatic radiative equilibrium(Chiang & Goldreich 1997; Dullemond et al. 2001). This provides strong observa-tional support for the physical assumptions underlying the flared-disk models.

The disk geometry plays an important role in the apparent strength of the PAHemission features. Comparison between observations and models shows that the


Fig. 1. Image of HD97048 in a narrow-band filter around the 8.6 μm PAH feature (Lagage

et al. 2006). The scale is the square-root of the flux to enhance the low-flux regions.

Courtesy of E. Pantin and P.-O. Lagage.

far-IR excess of Herbig Ae stars is connected to the degree of flaring of the dustdisk. The excess is produced by thermal emission of dust grains in the disk. Strongflaring leads to a large far-IR excess (Meeus et al. 2001; Dullemond & Dominik2004; Meijer et al. 2008; Acke et al. 2009). Dullemond et al. (2007) have modeledthe effect of dust sedimentation on the appearance of the disk and the strengthof the PAH features. Sedimentation makes the dust disk more flattened, andreduces the far-IR excess. However, the gas disk, containing the PAH molecules,remains flared. If dust sedimentation is the dominant process, the models predictstronger PAH emission in flattened disks, both because of an absolute increase inPAH luminosity and an increased feature-to-continuum contrast. Remarkably, theopposite trend is observed: several authors have shown that sources with a flareddust disk have stronger PAH features than those with flattened disks (Meeus et al.2001; Acke & van den Ancker 2004; Habart et al. 2004a; Keller et al. 2008; Ackeet al. 2010). In Figure 2, the increase of the feature-to-stellar luminosity ratio ofthe 6.2μm feature with increasing far-IR excess is shown for a sample of Herbig Ae


Fig. 2. The feature-to-stellar luminosity ratio of the 6.2 μm feature increases with in-

creasing 30 μm excess above the stellar photosphere (Acke et al. 2010). Flattened dust

disks are characterized by a low, and flared disks by a high far-IR excess.

stars. This correlation has been interpreted as a simple geometric effect: a flareddisk has a larger illuminated surface than a flattened disk. However, the linkbetween disk geometry and PAH emission strength is not necessarily a causal link.PAH molecules in the disk surface layer are an important contributor to the gasthermal balance due the photo-electric effect. Meeus et al. (2010) have shown thatinclusion of PAH molecules in different charge states can alter the gas temperatureconsiderably. The disk becomes warmer and the degree of flaring larger. PAHshence help to sculpture the disk structure (see Kamp, elsewhere in this volume).

A few objects with flattened dust disks (i.e. low far-IR excesses) are still strongPAH sources (van der Plas et al. 2008; Fedele et al. 2008; Verhoeff et al. 2010).In those targets, the scenario of dust sedimentation could be at work. The generaltrend, however, shows that sedimentation is not the dominating process in mostdisks. It indicates that either dust and gas remain well coupled (e.g. due to strongturbulent motions), or that the PAH molecules coagulate with dust grains andsettle along with the dust (Dullemond et al. 2007).

Because of the low detection rate of PAH emission in the lower-mass stars, thestatistics in this group of objects is too poor to perform a similar study on diskgeometry. Bouwman et al. (2008) note that the T Tauri stars without detectedhydrocarbon emission in their sample have a SED consistent with a flatteneddisk structure. Geers et al. (2006) mention that gaps in the inner dust diskneed to be taken into account as well. Such gaps may be formed due to the


Fig. 3. The 7.7 μm feature shifts to longer wavelengths with decreasing stellar effective

temperature.

gravitational interaction with a planetary system or a companion star. A gap withan extent of several to tens of AU reduces the thermal continuum emission of thedust at near- and mid-IR wavelengths. This enhances the feature-to-continuumcontrast. Moreover, the disk beyond the gap outer radius receives more directstellar photons. As a result, the disk becomes warmer, giving rise to an increasein scale height, and the surface area exposed to stellar radiation becomes largerthan in a gap-less disk. Next to the increased feature-to-continuum contrast, thisenhances the absolute PAH luminosity.

4 Chemistry and stellar radiation field

Peeters et al. (2002, and elsewhere in this volume) defined three classes of objectsbased on the peak positions and the relative contribution of the sub-bands of thePAH features in the 6–9μm range. Class A contains sources with blue 6.2 and7.7μm features. These are mostly general ISM sources, Hii regions, and reflectionnebulae, including those around Herbig Be stars. Class C sources are characterizedby red features. Very little objects belong to this Class: cool carbon-rich evolvedstars, red giant stars and a few T Tauri stars. The Herbig Ae stars belong toClass B (intermediate feature positions), together with most planetary nebulaeand a number of binary post-AGB stars. Sloan et al. (2005, 2008) note that the


PAH spectra of Herbig Ae stars are different from those of evolved stars belongingto Class B. Herbig Ae stars display a 7–9μm feature with an extended red wingbeyond 8.0μm. The authors therefore propose to create a new Class (dubbed B′)for the latter.

Boersma et al. (2008) show that the spatially resolved spectra of two Herbig Aestars exhibit Class A (ISM-like) characteristics far from the central star, andClass B characteristics closer in. The authors argue that all the emission em-anates from the surrounding nebula, and not from the disk. This shows thatchemical changes can occur even in the surrounding cloud.

In young stars, the Peeters classification is linked to the radiation field of thecentral star: the wavelength position of the 6.2 and 7.7μm features moves to thered with decreasing stellar effective temperature (Sloan et al. 2007; Keller et al.2008; Acke et al. 2010). In Figure 3, the Spitzer-IRS 6.5–10μm spectra of fiveyoung stars are shown. For clarity, the underlying continuum has been subtractedand the spectra are normalized to the peak flux of the 7.7μm feature. The figureillustrates the gradual shift of the feature position to shorter wavelengths with in-creasing stellar effective temperature, from Class C for the T Tauri star to Class Afor the Herbig Be star.

The interpretation of this shift is still a matter of debate. Sloan et al. (2005)suggested that the redshift of the main features indicates a higher aliphatic/aromatic (sp3/sp2) hybridization of the hydrocarbon molecules. This hypothe-sis is supported by the laboratory measurements of Pino et al. (2008). Acke et al.(2010) show that in Herbig Ae stars the redshift of the features, or equivalentlyTeff , is connected to an increase in relative strength of the features at 6.8 and7.2μm. The latter features are ascribed to aliphatic CH2/CH3 bending modes(e.g., Furton et al. 1999; Chiar et al. 2000; Dartois et al. 2007). Moreover,the strength of the 8.6μm feature, relative to the CC features at 6.2 and 7.7μm,decreases with feature redshift.

Rapacioli et al. (2005), Berne et al. (2007), Joblin et al. (2008), and Berneet al. (2009) follow a different approach (see also Berne et al., elsewhere in thisvolume). This group has investigated the spatially resolved IR spectra of photo-dissociation regions (PDRs) and PNe, and found that they could be decomposedinto a few fundamental basis spectra. Based on their fractional contribution tothe emission in different spatial regions of PDRs, the fundamental spectra areattributed to different families of hydrocarbon molecules/grains: ionized PAHs,neutral PAHs and very small grains (VSGs; possibly PAH clusters: see Rapacioliet al., elsewhere in this volume). Berne et al. (2009) fit the observed PAH spectraof a sample of young stars with the basis spectra. They find that the fractionof the emission ascribed by the authors to VSGs decreases with increasing UVluminosity of the central star. This correlation is very likely the same as the onementioned above between feature redshift and stellar effective temperature. TheVSG spectrum used by Berne et al. contains the reddest features in their spectralbasis. A reduction in the contributing fraction of the VSG spectrum is thereforeequivalent to a blueshift of the features. The link between UV luminosity and Teff

is also clear: cooler stars have in general weaker photospheric UV fields. However,


as opposed to the aliphatic/aromatic interpretation outlined above, Berne et al.suggest that an increased contribution of VSGs produces the position shift in thefeature.

A third possible explanation, which has not been explored sufficiently in theliterature, is that the PAH size and charge influence the spectral classification(Bauschlicher et al. 2008, 2009). The link between feature position and Teff couldthen reflect the excitation and charge state of the PAHs with respect to the stellarradiation field.

Keller et al. (2008) find that the flux ratios of the 12.7 and 11.3μm features,and of the 7.7 and 11.3μm features correlate with stellar effective temperature.The 12.7/11.3 ratio probes the importance of aromatic duo/trio CH modes withrespect to solo CH modes (Hony et al. 2001). The ratio is expected to be highin small PAHs, or PAHs with jagged edges. The features in the 6–9μm rangeare ascribed to ionized PAHs, and those at longer wavelengths to neutrals. The7.7/11.3 ratio is therefore interpreted as a measure for the ionization of the PAHs.Keller et al. (2008) conclude that PAH molecules around hotter stars have rougheredges and are more ionized, as a result of the stellar irradiation.

Finally, Visser et al. (2007) point out that the flux ratio of the 3.3 and 6.2μmfeatures is another strong probe of the PAH charge. The ratio increases by anorder of magnitude by going from ionized to neutral species. In the Herbig Ae/Besample of Acke & van den Ancker (2004), the detection rate of the 3.3μm feature,given the presence of the 6.2μm feature, is 44%: half of the PAH spectra aredominated by neutrals, the other half by ionized species. However, no link withstellar UV field or effective temperature is found.

5 Conclusions

The emitting hydrocarbon molecules in disks around young stars are different fromthose in the ISM, and their chemistry – aliphatic/aromatic hybridization ratio,clustering, size, and/or charge state – is connected to the stellar radiation field.There is an interplay between the disk structure and the strength of hydrocarbonemission. Targets with flared disks have brighter PAH features. In turn, thePAHs in the disk surface contribute to the gas thermal balance, shaping the diskmorphology.

References

Acke, B., & van den Ancker, M.E., 2004, A&A, 426, 151

Acke, B., Min, M., van den Ancker, M.E., et al., 2009, A&A, 502, L17








Bouwman, J., Henning, T., Hillenbrand, L.A., et al., 2008, ApJ, 683, 479

Cesarsky, D., Lequeux, J., Ryter, C., & Gerin, M., 2000, A&A, 354, L87

Chiang, E.I., & Goldreich, P., 1997, ApJ, 490, 368

Chiar, J.E., Tielens, A.G.G.M., Whittet, D.C.B., et al., 2000, ApJ, 537, 749

Dartois, E., Geballe, T.R., Pino, T., et al., 2007, A&A, 463, 635

de Graauw, T., Haser, L.N., Beintema, D.A., et al., 1996, A&A, 315, L49

Doucet, C., Habart, E., Pantin, E., et al., 2007, A&A, 470, 625

Dullemond, C.P., Dominik, C., & Natta, A., 2001, ApJ, 560, 957

Dullemond, C.P., & Dominik, C., 2004, A&A, 417, 159

Dullemond, C.P., Henning, T., Visser, R., et al., 2007, A&A, 473, 457

Fedele, D., van den Ancker, M.E., Acke, B., et al., 2008, A&A, 491, 809

Fedele, D., van den Ancker, M.E., Henning, T., et al., 2010, A&A, 510, A72

Furlan, E., Hartmann, L., Calvet, N., et al., 2006, ApJS, 165, 568

Furton, D.G., Laiho, J.W., & Witt, A.N., 1999, ApJ, 526, 752

Geers, V.C., Augereau, J., Pontoppidan, K.M., et al., 2006, A&A, 459, 545

Geers, V.C., 2007, Ph.D. Thesis, Leiden Observatory, Leiden University

Geers, V.C., van Dishoeck, E.F., Visser, R., et al., 2007, A&A, 476, 279

Geers, V.C., van Dishoeck, E.F., Pontoppidan, K.M., et al., 2009, A&A, 495, 837

Habart, E., Verstraete, L., Boulanger, F., et al., 2001, A&A, 373, 702

Habart, E., Natta, A., & Krugel, E., 2004a, A&A, 427, 179

Habart, E., Testi, L., Natta, A., & Carbillet, M., 2004b, ApJ, 614, L129

Habart, E., Natta, A., Testi, L., & Carbillet, M., 2006, A&A, 449, 1067

Herbig, G.H., 1960, ApJS, 4, 337


Houck, J.R., Roellig, T.L., van Cleve, J., et al., 2004, ApJS, 154, 18



Kessler, M.F., Steinz, J.A., Anderegg, M.E., et al., 1996, A&A, 315, L27

Lagage, P.O., Doucet, C., Pantin, E., et al., 2004, The Messenger, 117, 12

Lagage, P., Pel, J.W., Authier, M., et al., 2006, Science, 314, 621

Meeus, G., Waters, L.B.F.M., Bouwman, J., et al., 2001, A&A, 365, 476

Meeus, G., Pinte, C., Woitke, P., et al., 2010, A&A, 518, 124

Meijer, J., Dominik, C., de Koter, A., et al., 2008, A&A, 492, 451

Oliveira, I., Pontoppidan, K.M., Merın, B., et al., 2010, ApJ, 714, 778


Pino, T., Dartois, E., Cao, A., et al., 2008, A&A, 490, 665

Preibisch, T., Kim, Y., Favata, F., et al., 2005, ApJS, 160, 401

Quanz, S.P., Henning, T., Bouwman, J., et al., 2007, ApJ, 668, 359


Siebenmorgen, R., & Krugel, E., 2010, A&A, 511, A6


Sloan, G.C., Keller, L.D., Forrest, W.J., et al., 2005, ApJ, 632, 956


Sloan, G.C., 2008, in IAU Symposium, ed. S. Kwok & S. Sandford, 251, 191

Stelzer, B., Micela, G., Hamaguchi, K., & Schmitt, J.H.M.M., 2006, A&A, 457, 223

van Boekel, R., Waters, L.B.F.M., Dominik, C., et al., 2004, A&A, 418, 177

van der Plas, G., van den Ancker, M.E., Fedele, D., et al., 2008, A&A, 485, 487

Verhoeff, A.P., Min, M., Acke, B., et al., 2010, A&A, 516, A48

Visser, R., Geers, V.C., Dullemond, C.P., et al., 2007, A&A, 466, 229


EVOLUTION OF PAHS IN PROTOPLANETARY DISKS

I. Kamp1

Abstract. Depending on whom you ask, PAHs are either the smallestdust particles or the largest gas-phase molecules in space. Whetherreferred to as gas or dust, these PAHs can contain up to 20% of thetotal cosmic carbon abundance and as such also play an important rolein the carbon chemistry of protoplanetary disks. The interpretationof PAH bands is often a complex procedure involving not only gasphysics to determine their ionization stage and temperature, but alsoradiative transfer effects that can bury these bands in a strong thermalcontinuum from a population of larger dust particles.

PAHs are most readily seen in the spectral energy distributions(SEDs) of disks around Herbig AeBe stars where they are photopro-cessed by the stellar radiation field. Resolved images taken in the PAHbands confirm their origin in the flaring surfaces of circumstellar disks:if the SED is consistent with a flat disk structure (less illuminated),there is little or no evidence of PAH emission. The very low detec-tion rates in the disks around T Tauri stars often require an overalllower abundance of PAHs in these disk surface as compared to that inmolecular clouds.

In this review, I will adress three aspects of PAHs in protoplanetarydisks: (1) Do PAHs form in protoplanetary disks or do they originatefrom the precursor molecular cloud? (2) Is the presence of PAH featuresin SEDs a consequence of the disk structure or do PAHs in fact shapethe disk structure? (3) How can we use PAHs as tracers of processesin protoplanetary disks?

1 Introduction

Polycyclic aromatic hydrocarbons (PAHs) are considered amongst the smallestdust grains even though their sizes range from macromolecules with 10 carbonatoms, such as naphthalene to large complexes with more than 100 carbon atoms.

1 Kapteyn Astronomical Institute, University of Groningen, PO Box 9513, 2300 AVGroningen, The Netherlands



They are identified mainly through their near- and mid-IR bands in very differentastrophysical environments such as the interstellar medium (ISM), disks aroundyoung stars, and the circumstellar environment of AGB stars. I will focus in thefollowing on the role of PAHs in star formation and especially on their relevancein protoplanetary disks.

Protoplanetary disks are complex in nature, because they span a wide rangeof physical conditions, i.e. density and temperature (and irradiation), but alsoshow orders of magnitude different dynamical timescales between the inner andouter disk. Gas and dust in these disks are not necessarily coupled and co-spatial.And a number of processes that operate in protoplanetary disks such as dustcoagulation and settling, gas dispersal, and planet formation affect gas and dustoften in different ways.

Their nature positions PAHs in terms of their properties somewhere betweenlarge gas molecules and small dust grains. Due to their small size, they efficientlycouple to the widespread gas in protoplanetary disks and are thus able to escapesettling processes that affect the much larger dust grains. Due to their low ioniza-tion potentials, PAHs are easily ionized by the stellar UV photons and the ejectedelectrons can efficiently heat the surrounding gas (photoelectric heating). If theyare abundant (23% of carbon dust in the form of PAHs), they can also present animportant opacity source and absorb up to 40% of the total radiation (Habart et al.2004). Often the fraction of carbon dust in PAHs is assumed to be much smallerin disks, e.g. less than 10% (Geers et al. 2006; Dullemond et al. 2007a). PAHsare stochastically heated particles and undergo rapid and extreme temperaturefluctuations that lead to their emission in the aromatic infrared bands (AIBs).

I start out with a brief summary of PAH observations in protoplanetary disksand their detection frequency (Sect. 2). Section 3 provides a short overview of thePAH excitation mechanism and radiative transfer models. The next two sectionsthen discuss the role of PAHs in disks chemistry (Sect. 4) and physics (Sect. 5) inmore detail. The last section will then address three questions concerning PAHevolution in disks: (1) Do PAHs form in protoplanetary disks or do they originatefrom the precursor molecular cloud? (2) Is the presence of PAH features in SEDsa consequence of the disk structure or do PAHs in fact shape the disk structure?(3) How can we use PAHs as tracers of processes in protoplanetary disks?

2 Some observational facts

PAHs have been rarely observed in embedded low mass protostars. However, asthe sources evolve into class ii and class iii objects, detection rates increase andthere is a clear dependency on the UV radiation field of the central star. PAHsare observed in 57% of the Herbig disks (Acke et al. 2004), but only in 8% of theT Tauri disks (Geers et al. 2006). Various studies (Sloan et al. 2007; Boersmaet al. 2008; Keller et al. 2008) noted clear changes in the feature shape and shiftsof the central wavelength, which they attribute to signatures of PAH processingin the young star’s environment (see Acke, elsewhere this volume, for a moredetailed discussion). It seems that the wealth of observational data especially

I. Kamp: Evolution of PAHs in Protoplanetary Disks 273

from Spitzer available now for several objects (see below) warrants a much moredetailed understanding of the disk structure. For example, Fedele et al. (2008) findthe PAH emission coming from the upper gas surface in the disk around HD 101412;the PAHs are co-spatial with the [O i] 6300 A emission of the gas, which is generallythought to be a by-product of the photodissociation of OH in the UV irradiateddisk surface. On the other hand, the SED and near-IR visibilities show that thelarger dust grains are not directly exposed to the stellar radiation and must resideat lower heights in the disk. This lends support to the idea that PAHs stay wellmixed with the gas. Along the same lines, Verhoeff et al. (2010) find that the diskaround HD95881 consists of a thick puffed up inner rim and an outer region inwhich the gas still has a flaring structure, while larger dust grains have settled tothe midplane. Such systems are thought to be in a transitional stage between agas-rich flaring and gas-poor shadowed configuration.

3 PAH emission from disks

A detailed understanding of the PAH emission mechanisms is essential to judgethe diagnostic power and impact of PAHs on the disk structure. In the following,the problem is split into the excitation mechanism (ionization state, single versusmulti-photon excitation) and the radiative transfer in the presence of a strong dustcontinuum and competing features such as the silicate emission at 10 μm. Thecombination of these two aspects leads to differences in the spatial appearance ofthe various PAH emission features (see e.g. Fig. 1).

Studies of galactic emission have shown that PAHs contain up to a few per-cent of the total carbon dust mass (e.g. Pendleton et al. 1994); the typical PAHcarbon abundance found from galactic PDRs and ISM is [C/H]PAH ≈ 5 × 10−5

(Boulanger & Perault 1988; Habart et al. 2001; Desert et al. 1990; Dwek et al. 1997;Li & Draine 2001). For such an extreme case of PAH abundance in disks, Habartet al. (2004) find that PAHs reprocess ∼ 30% of the impinging stellar radiation(which is ∼0.3% of the total luminosity in the case of a flaring disk surface) in afiducial Herbig disk model (0.3 − 300 AU, 0.1 M�, M∗ = 2.4 M�, L∗ = 32 L�);since PAHs then present a significant fraction of the total dust opacity, their pres-ence has a visible effect on the overall SED (see their Figs. 1 and 3). Habart et al.(2004) find that ∼25% of the flux that PAHs absorb in the inner disk is consumedby their evaporation. Geers et al. (2006) found from an extensive comparison ofobserved PAH features (Spitzer spectra) and detailed radiative transfer disk mod-els that the typical PAH abundance in disks must be 10–100 times lower than thatin the ISM. Hence, the total dust opacity should be hardly affected by PAHs.

3.1 The excitation mechanism of PAHs

The intensity emitted in a specific AIB roughly scales with the incident FUV ra-diation provided that the charge state is held constant (Li & Lunine 2003; Habartet al. 2004). This directly reflects the single photon excitation mechanism (tran-siently heated grains) as opposed to the radiative equilibrium that holds for larger


Fig. 1. The SEDs of a T Tauri disk model (Model T1) and a Herbig disk model (Model

H1) at an evolutionary age of 106 yr (inclination of 45 degrees, d = 100 pc). The different

curves show the SED before sedimentation (solid), and after sedimentation (t = 106 yr)

for several values of the disk viscosity parameter: α = 10−2 (dotted) , α = 10−4 (dashed)

and α = 10−6 (dot-dashed). The figure is taken from Dullemond et al. (2007a).

grains. Hence, for PAHs one generally defines a temperature distribution functiondP/dT that describes the probability of finding a PAH in a particular tempera-ture interval. Due to the single photon excitation mechanism, this temperaturedistribution depends not on the strength, but on the color of the radiation field.The different PAH bands then require different excitation temperatures, with theshort wavelength bands (C-H, C-C stretching modes) requiring higher tempera-tures than the long wavelength bands (bending modes).

However, spinning of charged dust grains at thermal rates is an alternativeexcitation mechanism that works at much longer wavelengths Rafikov (2006) (seeDraine elsewhere this volume). This microwave emission would trace small dustgrains in general (VSGs and PAHs) and it has the potential to trace PAHs evendown to the midplane.

3.2 Feature-to-continuum ratios

Manske & Henning (1999) performed a two-dimensional radiative transfer calcu-lation for disk plus envelope systems. They include various grain populations,amongst them also the transiently heated VSGs and PAHs. One of their mainresults is that the PAH features become weaker with increasing mass fraction ofVSGs. However, it proved difficult to entirely surpress the PAHs in the SEDsfrom Herbig stars with disks and envelopes. The later models by Dullemond et al.(2007a) confirm that the VSGs compete with PAHs for the stellar UV photons,thus decreasing the PAH band strength. The SED on the other hand changes onlyslightly between 20 and 30 μm where VSGs produce a bump in the disk emission.

3.3 Effects of ionization, dehydrogenation and size

The various AIBs are attributed to vibrations of different bonds in the PAHmolecule and their strength changes with the PAH ionization, dehydrogenation


state, and size. The following paragraphs summarize our understanding of PAHemission based on results from Habart et al. (2004).

The 3.3 μm band is a C-H stretching mode, while the 6.2 and 7.7 μm bandsoriginate from C-C stretching modes. The C-H in-plane and out-of-plane bendingmodes cause the 8.6 μm band and the 11.3, 11.9, and 12.7 μm features respectively.

The strong effect that the ionization state (neutral or cation) has on the rel-ative PAH band intensities is known from theory and experiment (see referencesin Pauzat and Oomens, elsewhere in this volume). The 6.2 and 7.7 μm bandsare stronger in ionized PAHs; the 8.6 μm band is stronger in cations than in neu-trals. On the contrary, the 11.3, 11.9, 12.7 μm bands decrease for charged PAHs.Otherwise, PAH anion band strength is often between that of neutrals and cations.

With strong dehydrogenation, the C-H features would disappear. This is onlyrelevant for the smallest PAHs (NC ≤ 25) (Tielens et al. 1987; Allain et al. 1996).Theory predicts that the C-H features (3.3, 8.6, 11.3, 12.7) become weaker, whilethe C-C features (6.2, 7.7) become slightly stronger.

In addition to charge and hydrogenation state, the size of the PAH also impactsthe excitation state. Small PAHs favor the 3.3 μm band, while larger ones emitpreferentially at longer wavelengths. This leads to a degeneracy between chargeand size.

3.4 The spatial origin of various PAH features

Habart et al. (2006) spatially resolved the 3.3 μm PAH feature using NAOS-CONICA at the VLT (AO providing ∼ 0.1′′ resolution). The emission originatesfrom within 30 AU of the star and is significantly more extended than the adjacentcontinuum. This is in agreement with the theoretical models that find the higherexcitation temperatures required for the 3.3 μm feature mostly in the inner disk(r < 50 AU; Habart et al. 2004). Geers et al. (2007a), (2007b) also report that thedisk emission in some targets is more extended within the PAH bands than withinthe adjacent continuum. Just as a note of caution, the emission from VSGs canalso be more extended than that of large dust grains.

4 PAH chemistry in disks

As large gas molecules, PAHs take part in the chemical reaction networks of pro-toplanetary disks. They can undergo ionization (photoelectric effect), electronrecombination and attachment, charge exchange, photodissociation with C or Hloss, and H addition reactions. Trapping of PAHs onto dust grains or within ices(Gudipati & Allamandola 2003) are more relevant in the cold dense midplane thatdoes not contribute to the PAH emission features. Reaction rates with heavierelements can be comparable to those of H, H2, but often the abundances of thoseheavier elements are much lower. However, in some regions of protoplanetarydisks, where H is efficiently converted into water and/or OH, reactions with O andOH can play an important role in breaking PAHs down (Kress et al. 2010).


The convention often found in the literature (e.g. Le Page et al. 2001) is tocharacterize PAHs by the number of C atoms NC . The normal number of hydrogenatoms, N0

H is the number of hydrogen atoms when one hydrogen atom is attachedto each peripheral carbon atom which is bonded to exactly two other carbon atoms.In the following, PAH stands representative for any size PAH and the differentcharge states are indicated by suerscripts, e.g. PAH+. A recent review on PAHchemistry is given by Tielens (2008) and the formation of PAHs is discussed byCherchneff (this volume).

4.1 PAH hydrogenation and dissociation

The stability and hydrogenation of PAHs is mainly determined by the strength ofthe UV environment (λ < 2000 A) in which they reside (the ionization potentialof PAHs is typically IP > 6 eV, see Weingartner & Draine 2001). While photonsof 3–10 eV can lead to the loss of H (3.2 eV), C2H2 (4.2 eV), C (7.5 eV) andC2 (9.5 eV), X-ray photons are more energetic and can destroy entire PAHs withNC = 100 (Siebenmorgen & Krugel 2009).

Small PAHs (NC = 24) are easily destroyed in protoplanetary disks on shorttimescales (< 3 Myr). Larger PAHs survive more easily in the irradiation environ-ment of protoplanetary disks. PAHs with NC = 96 for example survive down to5 AU in the disk around a Herbig star, while smaller ones (NC = 50) are efficientlydestroyed out to a few 10 AU (see Fig. 7 of Visser et al. 2007). The weaker radi-ation field of T Tauri stars is not capable of destroying PAHs unless these starshave a significant UV excess. X-ray and EUV photons are so efficient that theydestroy PAHs wherever they can reach them (Siebenmorgen & Krugel 2009).

4.2 Charge distribution of PAHs in disks

The PAH charge state in disks generally has a layered structure with cationsbeing on top, followed by neutrals and anions towards the dense shielded midplane(Visser et al. 2007). Figure 2 shows that PAHs get easily double ionized (PAH2+)in the surfaces of disks around Herbig stars due to the strong stellar UV radiationfield (detailed model description in Sect. 5). This result depends strongly on thesize of the PAH. The disk model shown here uses circumcoronene (NC = 54,NH = 18, IP(1) = 6.243, IP(2) = 9.384, IP(3) = 12.524). Due to the weak UVradiation field of T Tauri stars, disks around low mass stars show mostly neutralPAHs and anions (∼ 50% of the entire population, Visser et al. 2007).

4.3 H2 formation on PAHs

Jonkheid et al. (2006) consider that for evolved disks (HD 141569), H2 might formmore efficiently in reactions with PAHs than on the dust grains in the disk (1 μm <a < 1 cm). Due to the lack of small grains, the total dust surface area availablefrom such a distribution is small compared to the typical ISM size distributionand the H2 formation rate depends directly on the total available surface area.


1 10 100r [AU]

0.0

0.1

0.2

0.3

0.4

0.5

0.6

z / r

AV

=1

AV=1

AV=10

0 2 4 6 8 10log χ

1 10 100r [AU]

0.0

0.1

0.2

0.3

0.4

0.5

0.6

z / r

-1 0 1 2 3<PAHcharge>

f(PAH)=0.01 f(PAH)=0.01

Fig. 2. Left: strength of the UV radiation field χ in the protoplanetary disk around a

typical Herbig star (M∗ = 2.5 M�, Teff = 10 000 K). The main parameters are: Mdisk =

0.025 M�, a surface density power law profile ε = 1.0, a grain size distribution between

amin = 0.05 μm and amax = 1 mm with a power law exponent of 3.5, and a dust-to-gas

mass ratio of 0.01, a PAH abundance fPAH of 1% relative to the ISM. χ is defined as the

integral of the radiation field between 912 and 2050 A normalized to that of the Draine

field (Draine & Bertoldi 1996). Right: charge distribution of PAHs. Note that regions

where PAHs carry multiple positive charges have low particle densities (see Fig. 3).

1 10 100r [AU]

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

z / r

1000

K 1000K

AV=1 AV=1AV=10

4 6 8 10 12 14log n

<H> [cm-3]

1 10 100r [AU]

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

z / r

1000

K

AV=1 AV=1AV=10

4 6 8 10 12 14log n

<H> [cm-3]

f(PAH)=0.01 f(PAH)=0.00001

Fig. 3. Density structure of a protoplanetary disk around a 2.5 M� Herbig star (Teff =

10 000 K). The main parameters are: Mdisk = 0.025 M�, a surface density power law

profile ε = 1.0, a grain size distribution between amin = 0.05 μm and amax = 1 mm with

a power law exponent of 3.5, and a dust-to-gas mass ratio of 0.01. The left figure shows

a model with a PAH abundance relative to the ISM of 0.01 and the right figure one of

10−5. The black dashed lines show the location of the AV = 1 and 10 surfaces, while the

red contour line indicates the location of the hot gas (Tgas > 1000 K).


5 The role of PAHs in disks physics

PAHs are not relevant for the dust energy balance, i.e. they do not present asignificant fraction of the dust opacity. However, as large molecules, they efficientlyabsorb UV radiation and get photoionized. The energetic electrons released in thisway efficiently heat the surrounding gas and determine its temperature. Thus thevertical disk structure – as set by the gas temperature – depends sensitively onthis heating source.

5.1 The vertical disk structure

The thermo-chemical disk code ProDiMo calculates the vertical hydrostatic diskstructure, the chemical composition, and gas & dust temperatures self-consistently.The basic physics and chemistry of the code is summarized in Woitke et al. (2009)while additions and new features are described in Kamp et al. (2010) (UV radiativetransfer and UV-photochemical rates) and Woitke et al. (in preparation) (UVpumping, PAH heating and ray tracing of gas lines).

The models shown here (van der Plas et al. in preparation) use a chemicalnetwork consisting of 10 elements, 76 species and 973 reactions, among them alsothe five ice species: CO, CO2, H2O, CH4, and NH3. The inner disk has “softedges”, i.e. soft density gradients at the inner and outer radius.

Figure 3 shows two models of a disk around a Herbig star that differ only intheir fractional PAH abundance. The inner disk wall is directly illuminated by thestar while the gas behind it is irradiated under a shallow angle; hence, the innerrim has a higher gas temperature than the disk behind it. Under the assumptionof vertical hydrostatic equilibrium, this translates into a “puffed-up” inner rim.If the rim has a substantial optical depth, it can cast a shadow on the materialbehind it and thus cause a “shadow”, i.e. a flat disk structure (Dullemond et al.2007b).

The amount of shadowing depends crucially on how efficient the gas is heatedby the stellar radiation field. The gas and dust temperatures in these models de-couple in the surface layers down to AV ∼ 1. For large gas heating efficiencies,it is thus more difficult to cast a shadow. One of the strongest heating processesin the disk surface is photoelectric heating by PAHs. Hence, the disk with verylittle PAHs has a much lower heating efficiency. The main heating processes inthis case are the photoelectric effect on dust grains and the H2 formation heating.The overall disk structure is cooler and hence much flatter (see e.g. the AV = 1contour lines). In addition, the inner rim now casts a very strong shadow beyond3 AU.

5.2 PAHs and dust settling

Dullemond et al. (2007a) have studied the impact of grain size sorting with heightdue to dust sedimentation and turbulence equilibrium. They find that the sedi-mentation enhances the dust features originating from the smallest grains (Fig. 1).


Even though PAHs are potentially destroyed in the irradiated disk surface, rapidturbulent mixing ensures a continuous supply from deeper layers.

As a consequence, the inclusion of sedimentation lowers the required PAHabundances to produce a certain feature-to-continuum ratio. The impact on theSED is twofold: (a) sedimentation supresses the continuum by letting small grainsthat are efficiently heated sink down (b) sedimentation lowers the optical depthin surface layers, thereby letting all UV flux be absorbed by PAHs in surface andthus enhancing the feature strength.

6 Conclusions

After an extensive discussion of the PAH chemistry, the role they play in diskphysics and their radiative transfer, this section returns to the initially raisedthree questions and tries to give an answer based on our current understanding.

6.1 Do PAHs form in protoplanetary disks?

The formation of PAHs requires high densities and temperatures (∼ 1000 K) andhigh abundances or organic molecules such as e.g. acetylene in a low UV radiationenvironment. Those conditions could likely exist directly behind the inner rim ofa protoplanetary disk. Using various initial abundances of C2H2 and CH4 andCO and (T, P ) combinations, Morgan et al. (1991) showed that PAH formationcan in fact be very efficient in certain temperature/pressure regimes of the earlysolar nebula. More recently, Woods & Willacy (2007) used a full chemical network(including the formation and destruction of C2H2 and CH4 and CO) to show thatin a stationary protoplanetary disk (no mixing), benzene can indeed form inside3 AU and abundances of other organic species are also high (see Woods, elsewherethis volume). Kress et al. (2010) define a soot line for disks, which is the dividingline between the location where carbon compounds readily condense into PAHs athigh gas temperatures and that where they rather stay in simpler molecules suchas CO, CO2, C2H2, CH4. The remaining question is how these in-situ formedPAHs are mixed to the large radii where they are observed (up to several 10 AU).

Alternatively, PAHs could form in shocks. Desch & Connolly (2002) invokeshocks for the melting of chondrules in the inner solar nebula. If the shocks areassociated with the global disk structure e.g. because of spiral density enhance-ments that travel at a lower pattern speed compared to the gas in which they areembedded, those shocks could be recurring on an orbital timescale. The conditions(e.g. temperatures and densities) could then be very similar to those in the outerenvelopes of AGB stars where PAHs are known to form efficiently (Cherchneffet al. 1992; Cherchneff elsewhere this volume).

Important for the question whether PAHs can survive from an earlier ISMphase is the destruction timescale. PAH destruction is mostly driven by UV andX-ray radiation leading to dehydrogenation and photodestruction of the carbonskeleton. In most cases, UV photons can break the weaker C-H bonds thus leadingto dehydrogenation. If PAHs absorb more than 21 eV in an interval shorter than


their cooling time, the carbon skeleton breaks up, thereby destroying the PAH(Guhathakurta & Draine 1989). These processes will be most relevant in thesurfaces of disks that are exposed to direct UV and X-ray radiation. Anotherimportant process is the collisional destruction of PAHs (involving H, O and OH)at high temperatures and pressure (T ∼ 1000 K, P ∼ 10−6 bar). Kress et al.(2010) find that the overall destruction timescale of PAHs is set by the timescaleof the most stable one, since high C2H2 abundances are used to convert smallerPAHs back into bigger ones.

Something to keep in mind is that the small dust in the form of VSGs andPAHs presents a very distinct grain population in the ISM. The fact that theydo not follow the general power law size distribution (Draine 2004) could indicatethat they experience very different processing as compared to the bulk silicate andgraphite dust. An alternative production mechanism could thus be a continuousreplenishment of PAHs from larger dust grains. Habart et al. (2004) suggestedan evaporation of icy grain mantles within the disk. Since PAHs are ubiquitousin the ISM, they may have been included into ices during the cold molecularcloud phase. Rafikov (2006) suggested a PAH production through fragmentationof larger grains; a similar process has been discussed by Berne et al. (2009) for theVSGs. This could fit into the more recent ideas that the dust grain size distributionin disks reflects the outcome of continuous coagulation and fragmentation processes(Birnstiel et al. 2010).

6.2 Do PAHs shape the protoplanetary disk structure?

Photoelectric heating of even small amounts of PAHs is generally more efficientin heating the gas than the rest of the entire dust population. As shown bymany authors (e.g. Kamp & Dullemond 2004; Jonkheid et al. 2004; Nomura &Millar 2005; Gorti & Hollenbach 2008), gas and dust temperature de-couple inthe disk surfaces, above AV ∼ 1, and the higher gas temperature determines thevertical scale height and hence flaring of the disk. The recently developed diskmodeling code ProDiMo solves self-consistently for the gas chemistry, energybalance and disk structure (Woitke et al. 2009; Kamp et al. 2010) also includingPAH chemistry and physics (Woitke et al. in preparation). A recent study ofProDiMo disk models around Herbig stars shows that reducing the fractionalabundance of PAHs in disks leads in general to much flatter disk structures andeven strong shadows (van der Plas et al. in preparation).

6.3 How can we use PAHs as tracers of processes in protoplanetary disks?

The ambiguity whether PAHs are very small dust grains or large gas moleculesmakes them universal tracers of both disks components and widens their diagnosticpotential.

As small dust grains, their emission can tell us if and where such small solidparticles survive the settling and coagulation processes that lead otherwise to a


rapid grain growth to mm- and cm-sized dust (e.g. Natta & Testi 2004). PAHs canalso be used to trace the radial disk structure up to ∼ 100 AU. Since these grainsare stochastically heated, their emission features are generally more extended thanthat of the continuum. As an example, Geers et al. (2007a) imaged IRS48 (pre-viously classified as M dwarf and recently re-classified as A dwarf) in the 8.6, 9,11.3, 11.9, 18.7 μm filters. The 18.7 μm image shows a gap with a radius of 30 AU,while the shorter PAH bands are all centered on source. Hence, they conclude thatthe gap cannot be entirely devoid in dust. This object maybe caught in a shortlived phase very much related to transitional disks such as HD141569. Anotherexample for deviations from a simple smooth disk structure observed in PAH bandemission are presented in Doucet et al. (2005) for the Herbig star HD97048.

As large molecules, PAHs are thought to remain well mixed with the gas inthe disk. As such, the PAH emission can also probe the physical conditions in thedisk surfaces such as flaring and irradiation. VISIR imaging of the disk aroundHD 97048 in the PAH bands has shown that the observed emission, flaring angleand vertical scale height are consistent with the predictions from passive irradi-ated hydrostatic equilibrium disk models (Lagage et al. 2006). This gives supportto the idea that PAHs rather belong to the gas than to the dust component ofdisks.

Along the same lines, Grady et al. (2005) show for a sample of 14 Herbig disksthat the visibility from STIS coronagraphic imaging (sensitive to 0.5′′–15′′ dis-tances from star, hence r ≥ 50 − 70 AU) is correlated with the strength of PAHfeatures (particularly the 6.2 μm band). They report a correlation with disk flar-ing, and an anticorrelation with dust settling and the absence of any correlationswith SED type, far-IR slope, mass accretion rate or strength of H2 emission. Thiscould suggest that there is initially a simple stratification with size, very smallgrains such as PAHs staying up high in the disk. During the disk evolution thesesmall grains should be photodestroyed on short timescales, leaving a rather flatstructure behind since photoelectric heating on large grains is generally less ef-ficient. Grady et al. (2005) also suggest that settling first occurs in the innerdisk, so we should observe some disks that are flat inside and still flaring outside(if sufficient sample size available - since this might be a very short evolutionaryphase), e.g. HD 163296, DM Tau (faint ring with dark inner disk), which are botharound 5 Myr. HD169142 (∼ 6 Myr) is an example, where the inner disk maybede-coupled from outer disk (Grady et al. 2007). The spatially resolved PAH emis-sion and the Meeus group I SED classification are only consistent if the inner diskis substantially flatter than the outer disk.

PAHs are also good diagnostics of the stellar radiation field as the relativeband strength indicate the fractional charging and typical sizes of PAHs whichis directly related to the strength of the stellar radiation field. The stronger theradiation field, the less small PAHs can survive (e.g. Visser et al. 2007). Thisseems to be also confirmed by an analysis of the mid-IR emission of a sample of12 Spitzer sources spread over the spectral ranges F to B (Berne et al. 2009).

This work made use of the ProDiMo code developed by P. Woitke, W.-F. Thi and I. Kamp.


References

Acke, B., van den Ancker, M.E., Dullemond, C.P., van Boekel, R., & Waters, L.B.F.M.,2004, A&A, 422, 621



Birnstiel, T., Ricci, L., Trotta, F., et al., 2010, A&A, 516, L14


Boulanger, F., & Perault, M., 1988, ApJ, 330, 964


Desch, S.J., & Connolly, Jr., H.C., 2002, Meteorit. Planet. Sci., 37, 183

Desert, F., Boulanger, F., & Puget, J.L., 1990, A&A, 237, 215

Doucet, C., Pantin, E., Lagage, P.O., et al., 2005, in Protostars and Planets V, 8127

Draine, B.T., 2004, in Astron. Soc. Pacific Conf. Ser., 309, Astrophysics of Dust, ed.A.N. Witt, G.C. Clayton, & B.T. Draine, 691

Draine, B.T., & Bertoldi, F., 1996, ApJ, 468, 269

Dullemond, C.P., Henning, T., Visser, R., et al., 2007a, A&A, 473, 457

Dullemond, C.P., Hollenbach, D., Kamp, I., & D’Alessio, P., 2007b, Protostars andPlanets V, 555

Dwek, E., Arendt, R.G., Fixsen, D.J., et al., 1997, ApJ, 475, 565

Fedele, D., van den Ancker, M.E., Acke, B., et al., 2008, A&A, 491, 809

Geers, V.C., Augereau, J., Pontoppidan, K.M., et al., 2006, A&A, 459, 545

Geers, V.C., Pontoppidan, K.M., van Dishoeck, E.F., et al., 2007a, A&A, 469, L35

Geers, V.C., van Dishoeck, E.F., Visser, R., et al., 2007b, A&A, 476, 279

Gorti, U., & Hollenbach, D., 2008, ApJ, 683, 287

Grady, C.A., Schneider, G., Hamaguchi, K., et al., 2007, ApJ, 665, 1391

Grady, C.A., Woodgate, B.E., Bowers, C.W., et al., 2005, ApJ, 630, 958


Guhathakurta, P., & Draine, B.T., 1989, ApJ, 345, 230

Habart, E., Natta, A., & Krugel, E., 2004, A&A, 427, 179

Habart, E., Natta, A., Testi, L., & Carbillet, M., 2006, A&A, 449, 1067


Jonkheid, B., Faas, F.G.A., van Zadelhoff, G.-J., & van Dishoeck, E.F., 2004, A&A, 428,511

Jonkheid, B., Kamp, I., Augereau, J., & van Dishoeck, E.F., 2006, A&A, 453, 163

Kamp, I., & Dullemond, C.P., 2004, ApJ, 615, 991

Kamp, I., Tilling, I., Woitke, P., Thi, W., & Hogerheijde, M., 2010, A&A, 510, A260000


Kress, M.E., Tielens, A.G.G.M., & Frenklach, M., 2010, Adv. Space Res., 46, 44

Lagage, P., Doucet, C., Pantin, E., et al., 2006, Science, 314, 621

Le Page, V., Snow, T.P., & Bierbaum, V.M., 2001, ApJS, 132, 233

Li, A., & Draine, B.T., 2001, ApJ, 554, 778

Li, A., & Lunine, J.I., 2003, ApJ, 594, 987


Manske, V., & Henning, T., 1999, A&A, 349, 907

Morgan, Jr., W.A., Feigelson, E.D., Wang, H., & Frenklach, M., 1991, Science, 252, 109

Natta, A., & Testi, L., 2004, in Astron. Soc. Pacific Conf. Ser., 323, Star Formation in theInterstellar Medium: In Honor of David Hollenbach, ed. D. Johnstone, F.C. Adams,D.N.C. Lin, D.A. Neufeeld, & E.C. Ostriker , 279

Nomura, H., & Millar, T.J., 2005, A&A, 438, 923

Pendleton, Y.J., Sandford, S.A., Allamandola, L.J., Tielens, A.G.G.M., & Sellgren, K.,1994, ApJ, 437, 683

Rafikov, R.R., 2006, ApJ, 646, 288

Siebenmorgen, R., & Krugel, E., 2009 [ArXiv e-prints]



Tielens, A.G.G.M., Seab, C.G., Hollenbach, D.J., & McKee, C.F., 1987, ApJ, 319, L109

Verhoeff, A.P., Min, M., Acke, B., et al., 2010, A&A, 516, A48


Weingartner, J.C., & Draine, B.T., 2001, ApJS, 134, 263

Woitke, P., Kamp, I., & Thi, W.-F., 2009, A&A, 501, 383

Woods, P.M., & Willacy, K., 2007, ApJ, 655, L49


PAH IN VECTORIZED THREE DIMENSIONAL MONTECARLO DUST RADIATIVE TRANSFER MODELS

R. Siebenmorgen1, F. Heymann2 and E. Krugel3

Abstract. We present a Monte Carlo (MC) radiative transfer code forcomplex three dimensional dust distributions and include transientlyheated PAH. The correctness of the code is confirmed by comparisonwith benchmark results. The method makes use of the parallelizationcapabilities of modern vectorized computing units like graphic cards.The computational speed grows linearly with the number of graphicalprocessing units (GPU). On a conventional desktop PC, our code isup to a factor 100 faster when compared to other MC algorithms. Asan example, we compute the dust emission of proto-planetary disks.We simulate how a mid-IR instrument mounted at a future 42 m ELTwill detect such disks. Two cases are distinguished: a homogeneousdisk and a disk with an outward migrating planet, producing a gapand a spiral density wave. We find that the resulting mid-IR spectraof both disks are almost identical. However, they can be distinguishedat those wavelengths by coronographic, dual-band imaging. Finally,the emission of PAHs exposed to different radiation fields is computed.We demonstrate that PAH emission depends not only on the strengthbut also strongly on the hardness of the radiation, a fact which hasoften been neglected in previous models. We find that hard photons(> 20 eV) easily dissociate all PAHs in the disks of T Tauri stars. Toexplain the low, but not negligible detection rate (< 10%) of PAHs inT Tau disks, we suggest that turbulent motions act as a possible pathfor PAH survival.

1 Dust model

In the dust model we consider: large (60 A < a < 0.2 − 0.3 μm) silicate (Draine2003) and amorphous carbon (Zubko et al. 2004) grains and small graphite

1 ESO, Karl-Schwarzschild-Str. 2, 85748 Garching b. Munchen, Germany2 Astronomisches Institut, Ruhr–Universitat Bochum, Universitatsstra. 150, 44801 Bochum,Germany and ESO3 MPI fur Radioastronomie, Auf dem Hugel 69, Postfach 2024, 53010 Bonn, Germany



(5 < a < 80 A) grains. For both, we apply a power law size distribution: n(a) ∝a−3.5 and absorption and scattering cross-sections are computed with Mie theory.In addition we include PAHs with 30 and 200 C atoms. For the absorption crosssection (Draine 2010; Malloci 2010; Verstraete 2010) we follow at photon energiesbetween 1.7–15eV Malloci et al. (2007, their Fig. 4). At low frequency, we takea cross section cut-off given as an average of neutral (Schutte et al. 1993) andionized (Salama et al. 1996) species. For hard photons, beyond the 17 eV band,we scale the photo-absorption cross section of PAH up to the keV region to similarsized graphite particles. By computing cross sections above 100 eV, we consideran approximation of kinetic energy losses (Dwek & Smith (1996) and apply it toall dust particles. With the advent of ISO and Spitzer more PAH emission fea-tures and more details of their band structures are detected (Tielens 2008). Weconsider 17 emission bands and take, for simplicity and as suggested by Boulangeret al. (1998) and Siebenmorgen et al. (1998), Lorentzian profiles. Parameters arelisted in Siebenmorgen & Krugel (2001) and calibrated using mid-IR spectra ofstarburst nuclei. For starburst galaxies, a SED library is computed (Siebenmorgenet al. 2007) and those models provide good fits to the SED of local galaxies and toPAH detected at high red shifts (z ≈ 3, Efstathiou & Siebenmorgen 2009). In themodel, we use dust abundances of [X]/[H] (ppm) of: 31 for [Si], 150 [amorphousC], 50 [graphite] and 30 [PAH], respectively; which is in agreement with cosmicabundance constrains (Asplund et al. 2009). We are in the process of upgradingthe model to be consistent with the polarisation of the ISM (Voshchinnikov 2004).

2 Monte Carlo method

We compute the radiative transfer with a Monte Carlo (MC) technique, whichallows to handle complicated geometries, by following the flight path of manyrandom photons. The basic ideas of our method go back to Lucy (1999) andBjorkmann & Wood (2001). They are described and extended to the treatment oftransiently heated PAHs in Krugel (2006). The space is partitioned into cubes and,where a finer grid is needed, the cubes are further divided into subcubes (cells).The star emits photon packages of constant energy ε, but different frequencies. Apackage entering on its flight path a cell may be absorbed there or scattered. Theprobability for such an interaction is given by the extinction optical depth alongthe path within the cell. When the package is scattered, it only changes directiondetermined in a probabilistic manner by the phase function. When it is absorbed,a new package of the same energy, but usually different frequency νnew is emittedfrom the spot of absorption. The emission is isotropic. Each absorption eventraises the energy of the cell by ε, and accordingly its temperature.

We were able to increase the computational speed of the code by up to twoorders of magnitude by parallelizing it, i.e. by calculating the flight paths ofa hundred or more photons simultaneously. This allows the treatment of com-plex geometries which were so far prohibited because of their excessive com-puter demands. To achieve parallelization, we had to slightly modify the originalcode. The latter calculates the new frequency νnew of a package that is emitted

R. Siebenmorgen et al.: PAH and Radiative Transfer 287

after absorption using the cell temperature before absorption, but applies a cor-rection for the new temperature due to Bjorkmann & Wood (2001). We nowomit this correction or in, mathematical terms, use Equation (11.67) instead ofEquation (11.65) in Krugel (2006). As expected and borne out in tests, thecorrection may be neglected when the number of photons absorbed in a cell isnot small. The vectorized MC code was developed during the PhD project ofHeymann (2010). It was verified against benchmarks provided for stellar heateddust spheres by Ivezic et al. (1997), the ray tracing code by Krugel (2006) andaxi–symmetrical disks by Pascucci et al. (2004). Generally, the computed SEDagrees with the benchmarks to within a few percent. One example of such com-parisons is shown in Figure 1 for a dust sphere with PAHs. The factor by whichparallelization speeds up the computation scales almost linearly with the numberof graphical processing units. We point out that particular attention had to begiven to the choice of the random number generator where we chose the MerseneTwister algorithm (Matsumoto & Nishimura 1998).

4 6 8 10 12 14wavelength (μm)

1

10

flux

(Jy)

K06MC

Fig. 1. Comparison of a SED of a stellar heated dust sphere of constant density and

visual extinction to the star of AV = 10 mag computed with a ray tracing method, as

described in Krugel (2006), and the MC treatment of this work; both methods agree to

within a few %.

3 Detection of proto-planetary disk structures

Hydrodynamical simulations of proto–planetary disks with an orbiting planet showparticular disk features (Masset et al. 2006). We present a three dimensionalapplication of the MC code where a disk heated by a solar type star has a dustdensity which follows the one as given by Pascucci et al., and has in addition agap between 3 – 8 AU and a low density spiral structure. The optical depth alongthe midplane from 0.2 to 75AU is 10mag. We simulate if such disk structures canbe resolved at a distance of 50 pc with a mid-IR instrument (Brandl et al. 2010)mounted at a future 42 m extreme large telescope (ELT); a project under study byESO. The point spread function (PSF) of the ELT has resolution of 50milliarcsecat 10 μm. In order to improve the contrast between star and disk, the instrument


0.2 0.4 0.6 0.8size (’’)

10

100

flux

(μJy

)

3σ/1h

spiral

gap

disk with PAHdisk without PAHdisk without spiral

PSF

Fig. 2. Image at 11.3 μm (left) together with flux profiles along the major axis (right).

The 3σ detection limit (dotted) and PSF (dash-dotted) is indicated. For the disk with

(magenta) or without (black) PAH the gap and the spiral structure is well preserved. A

homogeneous disk including PAHs is shown in dased for comparison.

will provide coronographic and dual band imaging modes. We choose band passesat 11.3 and 10 μm. The 11.3 μm emission is shown in Figure 2 together with theflux profile along the major axis. Profiles of such a disk with and without PAHsand that of a homogeneous disk are computed. The 3σ detection limit after 1hintegration is given assuming background limited performance of the instrument.The models predict that the detailed structure of such a proto-planetary disk canbe well detected.

Fig. 3. The temperature distribution P (T ) of a PAH with 100 C atoms exposed to mono–

chromatic radiation with hν = 3.8, 13.6, 50, 100, 300 eV and 1 keV in a constant heating

bath of flux F = 104 erg s−1 cm−2 (Siebenmorgen & Krugel 2010). The shaded area

marks temperatures above the sublimation temperature of graphite.

R. Siebenmorgen et al.: PAH and Radiative Transfer 289

4 Destruction and survival of PAH in T Tauri disks

Despite the fact that from the stellar heating one would expect to detect PAH inthe disks of T Tauri, one find them rather seldom, in less than 10% of T Tauri(Geers et al. 2006). In order to explain this fact we present a simplified scheme toestimate the location from the primary heating source at which the PAH moleculesbecome photo-stable (Siebenmorgen & Krugel 2010). T Tauri stars have besidephotospheric emission also a far ultraviolet (FUV), an extreme ultraviolet (EUV)and an X–ray component with a fractional luminosity of about 1%, 0.1% and0.025%, respectively. Such hard photons are very efficient in dissociating PAHs.The temperature distribution, P (T ), of PAHs after hard photon absorption isshown in Figure 3. It demonstrates that P (T ) depends strongly on the hardnessand spectral shape of the exciting radiation field, a fact which is often neglected incomputations of the PAH emission. After photon absorption, a highly vibrationallyexcited PAH may relax through emission of IR photons or, if sufficiently excited,lose atoms (Omont 1986 & Tielens 2005) for a textbook description). We findthat hard photons (EUV and X-ray) would destroy all PAHs in the disk of TTauri stars; whereas soft photons with energies < 20 eV dissociate PAHs only upto short distances from the star (1–2 AU). As a possible path for PAH-survivalturbulent vertical motions are suggested. They can replenish or remove PAHsfrom the reach of hard photons. In our treatment the presence of gas is consideredwhich is ionized at the top of the disk and neutral at lower levels. A view of thescheme is shown as vertical cut along the midplane in Figure 4.

z

midplaner

l

2

0

τ=1

4

light from star

αg

extinction layer

z 0

τ

motionsturbulent

Fig. 4. Of each radiation components (photosphere, FUV, EUV and X-rays) of the

fiducial T Tauri star about 90% is absorbed in what we call extinction layer. The optical

depth from its bottom to the star is one and in the vertical direction equal to the grazing

angle αg. The height of its lower boundary z0 declines with the radius, but its geometrical

thickness is rather constant ( ≈ 0.5 H, for details see Siebenmorgen & Krugel 2010).

Vertical motions may either remove PAH from the extinction layer quickly enough for

survival or replenish PAH from below.


References

Asplund, M., Grevesse, N., Sauval, A.J., & Scott, P., 2009, ARA&A, 47, 481

Bjorkman, J.E., & Wood, K., 2001, ApJ, 554, 615

Boulanger, F., Boisssel, P., Cesarsky, D., & Ryter, C., 1998, A&A, 339, 194

Brandl B., et al., 2010, SPIE, 7735-86, in press

Draine, B.T., 2003, ApJ, 598, 1026

Draine, B.T., 2011, this volume

Dwek, E., & Smith, R.K., 1996, ApJ, 459, 686

Efstathiou, A., & Siebenmorgen, R., 2009, A&A, 502, 541

Geers, V.C., Augereau, J.-C., Pontoppidan, K.M., et al., 2006, A&A, 459, 545

Heymann, F., 2010, Ph.D. Thesis, University of Bochum

Ivezic, Z., Groenewegen, M.A.T., Menshchikov, A., et al., 1997, MNRAS, 291, 121

Krugel, E., 2006, An introduction to the Physics of Interstellar Dust (IoP)

Lucy, L.B., 1999, A&A, 344, 282


Malloci, G., 2011, this volume

Masset, F.S., Morbidelli, A., Crida, A., & Ferreira, J., 2006, ApJ, 642, 478

Matsumoto, M., & Nishimura, T., 1998, ACM Trans. Model. Comp. Simul., 8, 330

Micelotta, E.R., Jones, A.P., & Tielens, A.G.G.M., 2010, A&A, 510, A36

Omont, A., 1986, A&A, 166, 159

Pascucci, I., Wolf, S., Steinacker, J., et al., 2004, A&A, 417, 793

Salama, F., Bakes, E.L.O., Allamandola, L.J., & Tielens, A.G.G.M., 1996, ApJ, 458, 621


Siebenmorgen, R., Natta, A., Krugel, E., & Prusti, T., 1998, A&A, 339, 134

Siebenmorgen, R., Krugel, E., & Laureijs, L., 2001, A&A, 377, 735

Siebenmorgen, R., & Krugel, E., 2007, A&A, 461, 445

Siebenmorgen, R., & Krugel, E., 2010, A&A, 511, A6

Tielens, A.G.G.M., 2005, The Physics and Chemistry of the Interstellar Medium(Cambridge Univ. Press)


Verstraete, L., 2011, this volume

Voshchinnikov, N.V., 2004, Optics of cosmic dust I, ASPR, 12, 1


PAHs and Carbonaceous

Grains & Solar System Materials


FROM PAHS TO SOLID CARBON

C. Jager1, H. Mutschke2, T. Henning3 and F. Huisken1

Abstract. Carbonaceous grains represent a major component of cosmicdust. The review will summarize new results in laboratory investiga-tions of carbonaceous dust components. The nanometer-sized carbonparticles are supposed to represent a blend of differently structuredcarbon including graphitic, diamond-like, fullerene-like and chain-likecomponents on a subnanometer or nanometer scale. Recent modelsused to explain the structure of gas-phase condensed carbon nano-particles are discussed. Possible formation pathways of carbonaceousgrains from molecular components and clusters and the role of poly-cyclic aromatic hydrocarbons (PAHs) and fullerenes are disclosed.

1 Introduction

Carbonaceous grains belong to the most abundant cosmic dust components. Cos-mic dust absorbs and scatters stellar light and radiates the energy in the IR andmillimeter range. Very small dust grains influence the heating rate of gas in the in-terstellar medium (ISM) by the photoelectric effect. In addition, surfaces of dustgrains are efficient factories for the formation of molecular hydrogen and largerorganic molecules. In the diffuse ISM, grain shattering and sputtering processescan erode and fragment grains into smaller units and therefore grains can act asprecursors of organic molecules. Carbon and hydrogen atoms constitute the basisof a rich organic chemistry in the solar system and in the interstellar medium andfinally for life on earth and, perhaps, on other planets. However, the problem withthe carbon dust component is the huge diversity of allotropes and structures thatcan be mixed up in one grain. Due to the processing of the dust grains in theinterstellar medium, the composition and structure is constantly modified. Large

1 Max-Planck-Institute for Astronomy, Heidelberg and Institute of Solid State Physics,Friedrich Schiller University, Helmholtzweg 3, 07743 Jena, Germany2 Astrophysical Institute and University Observatory, Friedrich Schiller University,Schillergasschen 2-3, 07745 Jena, Germany3 Max-Planck-Institute for Astronomy, Konigstuhl 17, 69117 Heidelberg, Germany



PAH molecules are involved in these processes both as precursors of condensationand products of grain disintegration. Consequently, the spectral properties of cos-mic carbon grains, which are strongly related to the structure and composition ofthe carbonaceous materials, are changing considerably from circumstellar to in-terstellar to protoplanetary environments. In order to “understand” the spectralsignature of the cosmic carbon component, one has to study the spectral propertiesof differently structured carbon materials in the laboratory.

2 Solid forms of carbon

For carbon, the possibilities to form solid structures are extremely manifold be-cause of its ability to occur in different hybridization states. Carbon exist indifferent allotropes including diamond, graphite, polyynes and fullerenes. Furthersolid forms of carbon are graphene and nanotubes, new forms of solid carbon,which are based on graphitic or fullerene structures.

In the diamond lattice, the tetrahedrally coordinated carbon atoms are cova-lently bound to each other via four strong σ bonds. Consequently, diamond isan insulator having a large band gap of 5.5 eV. In graphite, the three sp2 hybridorbitals have a trigonal planar symmetry and can form σ bonds with the adjacentcarbon atoms whereas the nonhybridized p electron is able to form an additional πbond. This results in infinite layers of condensed benzene rings with delocalized πelectrons between them. Therefore, graphite represents a zero band gap material.

In polyynes, the carbon atoms are sp hybridized forming two sp hybrid orbitalswith two remaining p orbitals. The sp hybridized carbon atoms build up chainswith alternating triple bonds. Polyyne molecules or radicals are found to exist inthe ISM. These chain-like structures form in gas-phase condensation experimentsand they are supposed to be precursors of fullerenes or fullerene-like carbon grains(Ravagnan et al. 2002).

Apart from the distinct hybridization states, there exist mixed hybridizationstates which build up curved structures. In the perfect form, they are closed-shell molecules consisting of only carbon, called fullerenes (the fourth allotropeof carbon). They can be produced in gas-phase condensation reactions such asby laser ablation or arc discharge of carbon electrodes in helium gas atmosphere(Kroto et al. 1985; Kratschmer et al. 1990). The curvature is achieved by theincorporation of pentagons and heptagons. The deviation of the planar structureresults in a higher s-character of the π orbital and a mixed hybridization state(Haddon 1993). C60, C70, or smaller fullerenes are not “superaromatic” as theytend to avoid double bonds in the pentagonal rings. C60 behaves like an electrondeficient alkene and reacts readily with electron-rich species. Many fullerenes havenow been detected in carbon soot, uncovered by electron microscopy, includingvery small ones such as C20 or C34.

Carbon nanotubes (CNTs), the fifth form of solid carbon, can exist as single-walled or multi-walled species with several nested cylinders. Multi-walled formscan be produced in gas-phase condensations together with other particulate carbon

C. Jager et al.: From PAHs to Solid Carbon 295

forms (Iijima 1991). Single walled CNTs up to lengths of about 1 μm can beproduced using catalysts (Harris 2009). A single-wall CNT can be imagined asa graphene sheet rolled at a certain chiral angle. CNTs can grow up to lengthsover 1 μm and with a range of diameters from 1 nm (for single walled) to around50 nm for multi walled ones. The electronic properties vary from metallic tosemiconducting with diameter and chirality.

Since 2004, a sixth form of carbon has been defined that can be consideredas a single layer of graphite called “graphene”. The material has been producedby mechanical exfoliation of graphite (Geim & Novoselov 2007). Other methods,developed to produce graphene, are SiC and hydrazine reduction, or chemicalvapor deposition. One can consider graphene as an infinite or very large PAHpartly hydrogen deficient. The largest molecule chemically synthesized in thelaboratory is C222H42 with a mass of around 2700 u (Wu et al. 2007). In gas-phase condensation processes, even larger molecules can be produced, however, tomake them visible with analytical methods is rather difficult.

The very large graphene species have to be stabilized either by a substrate or bya scaffold. Graphene can be transformed into a hydrogenated form (graphane) byexposing the sheets to a cold hydrogen plasma (Savchenko 2009; Elias et al. 2009).The material transforms from a highly conductive material into an insulator. Thisprocess is reversible and can be reversed by annealing in argon. Density-functional-based tight-binding level calculations have shown that the HOMO-LUMO gaps ofgraphene nanoflakes vary from zero to gap energies typical for insulators (Kucet al. 2010).

3 Structure of gas-phase condensed carbon nano-particles

It is generally assumed that the major fraction of the primary cosmic carbona-ceous material is formed as nanometer-sized particles via gas-phase condensationin circumstellar envelopes of carbon-rich asymptotic giant branch (AGB) stars(Henning & Salama 1998). In order to understand the cosmic condensation ofcarbonaceous grains, gas-phase condensations of carbonaceous grains have to besimulated in the laboratory. Dedicated analytical methods such as high-resolutiontransmission electron microscopy (HRTEM) and UV/Vis/IR spectroscopy help tounderstand the structure and composition of condensed carbon grains. Here, wepresent a brief overview on the structural characteristics of gas-phase condensednano-particles produced from carbon and hydrogen.

The structural complexity of pure carbon or carbon mixed with hydrogen canbe huge in solid carbon. Differently hybridized carbon (sp, sp2, and sp3) andmixed hybridization states may occur in different ratios in one grain. The struc-ture of particulate carbon materials can be characterized on different length scalesdescribing the short-, medium-, and long-range order of the material. The long-range order describes the size and shape of the primary particles and their agglom-eration state. The medium-range order characterizes the arrangement, shape andsize of the structural subunits and can be derived by HRTEM. So-called basic


Fig. 1. HRTEM image of a gas-phase condensed particle having a paracrystalline struc-

ture. The chemical formula (right) shows one possible amorphous carbon structure which

can be considered as a paracrystalline graphene layers.

structural units (BSUs) consisting of small subnanometer- or nanometer-sizedgraphitic crystallites which can be randomly or concentrically arranged in a grainhave been already discussed in the sixties (Heidenreich et al. 1968). However,plenty of HRTEM studies of condensed particles show the presence of individualgraphene layers that are concentrically arranged. This structure can be describedby the more recent paracrystalline model of soot grains (Hess & Herd 1993) whichcan be characterized by a layer structure of perturbed graphene layers. At present,this concept is the most widely accepted microstructural model for soot.

The hybridization state of the carbon, the nature of the bonds between car-bon atoms as well as the incorporation of hydrogen or hetero atoms in the carbongrains can be specified by the short-range order of the material. Analytical meth-ods like electron energy loss, IR, or nuclear magnetic resonance spectroscopy haveto be applied to obtain the information on the short range order (Schnaiter et al.1998; Jager et al. 1999). The short- and the medium-range order of carbon grainscan be well described by the terms “amorphous carbon” (AC) or “hydrogenatedamorphous carbon” (HAC) (see Fig. 1). HACs are frequently used to describe thecarbonaceous dust component in astrophysical studies (Jones et al. 1990; Duley1994; Pendleton & Allamandola 2002; Dartois et al. 2010). Both terms, “HACs”and ACs, are misleading since they describe groups of carbonaceous materials withdifferent C/H ratios, varying ratios of sp3, sp2, and sp hybridized carbon, and gap


energies between 0.5 and about 4.0 eV. They represent three-dimensional struc-tures applicable for the characterization of graphitic or paracrystalline structuralunits in grains. In AC and HAC, the carbonaceous structure can be characterizedby aromatic islands of different sizes, which are linked by aliphatic structures in-cluding sp, sp2, and sp3 carbon atoms. However, there is no clear discriminationbetween AC and HAC on the basis of the sizes of aromatic and aliphatic units andon their H/C ratios. In astrophysical studies, the term HAC is preferentially usedfor carbon materials with very small aromatic units (2–8 rings Jones et al. 1990).Oxygen in the form of ether bridges (-C-O-C-) or carbonyl groups (-C=O) mayalso be included. However, the amount of oxygen incorporated into a solid hydro-carbon material present in the diffuse interstellar medium is assumed to be low(Pendleton & Allamandola 2002). According to Robertson (1986), the aromaticas well as the aliphatic structures in HACs tend to cluster. Another acronym forsimilar materials were “quenched carbonaceous composites” (Sakata et al. 1993;Wada et al. 1999). The application of natural coals as cosmic dust analogs hasbeen suggested by Papoular et al. (1993).

The spectral properties of HACs in the UV/Vis are dominated by electronictransitions. Whereas the σ − σ* transitions are located in the far UV, the π − π∗transitions occur at wavelengths between 190 and 270 nm. Robertson & O’ Reilly(1987) claimed that the size of the aromatic units determine the gap energy andthe position of the π − π* transitions. The gap energy Eg can be derived fromthe optical spectra by means of the Tauc relation

√ε2E = B(E − Eg) (Tauc

et al. 1966). According to the relation Eg = 2βM−1/2 derived by Robertson &O’Reilly (1987), where β represents a quantum chemically estimated overlappingenergy between two adjacent pz orbitals that has an empirical value of 2.9 eV,the optical gap energy of soot is related to the number M of condensed rings inthe graphene layers. However, nowadays it is widely accepted that bent graphenelayers can be incorporated into particles. For mixed hybridization states, theelectronic transitions are shifted to higher energies proven by using sophisticatedspectral measurements such as electron energy loss or near edge X-ray absorptionspectroscopy (Ajayan et al. 1993; Ravagnan et al. 2006; Jager et al. 2010).

In the IR range, the functional groups included in the HAC structure can causevibrational bands in absorption or emission. In addition, the sizes of aromaticsubunits also influence the IR spectral properties due to continuous absorptionsby free charge carriers. In particular, the sizes of aromatic subunits affect theconductivity of the material which has a strong influence on the absorptivity inthe range between 100μm and 1 mm.

A special group of gas-phase condensed grains are onion-like or fullerene-likegrains including nested buckyonions. Faceted polyhedral onions and onions withnot perfectly closed shells are formed during the condensation of carbon vapor athigh temperature, for example, in the arc discharge. Completely spherical onionsare non-equilibrium structures and self-assemble only under irradiation of graphiticor amorphous carbon precursors (Ugarte 1992; Banhart 1999). Nowadays, defec-tive carbon onions were discussed to be possible carriers of different observationalbands such as the 217 nm bump (Tomita et al. 2002; Chhowalla et al. 2003).


4 Soot formation

4.1 Soot formation in astrophysical environments

Carbon grain condensation in astrophysical environments has been rarely discussedin the astrophysical literature. Gail & Sedlmayr (1985) have applied classicalnucleation theory to study the carbon grain formation in stationary, sphericalexpanding winds. They calculated condensation temperatures of dust between1260 and 1320 K for optically thin and between 1350 and 1430 K for opticallythick shells. The final grain radii were found to be small but very close to thetheoretically possible maximum if all condensable material is included in grains.Lodders & Fegley (1999) modeled dust condensation in carbon-rich AGB starsusing thermodynamic equilibrium calculations and found that the condensationtemperature of graphite is pressure independent but sensitive to the C/O ratio inthe circumstellar shells.

Detailed kinetic models for the formation of PAHs from acetylene in AGB starshave been presented by Frenklach & Feigelson (1989), Cherchneff et al. (1992),Allain et al. (1997), Cherchneff (2010). The authors determined a narrow tem-perature range of 900 to 1100K for the formation of PAHs in circumstellar en-vironments strongly dependent on the basic set of conditions such as mass loss,wind velocity, and gas density. Frenklach & Feigelson (1989) found a yield of 1.3%with respect to the amount of carbon initially present in acetylene for the mostfavorable case (low effective temperature of 1500 K, a mass loss rate as high as10−4 M� per year and a wind velocity of 10−2 km s−1). Cherchneff et al. (1992)obtained a maximum yield of 5×10−5. Cherchneff & Cau (1999) reconsideredthe modeling of the PAH and carbon dust formation in carbon-rich AGB starsand applied a physico-chemical model, which describes the periodically shockedgas in the circumstellar shells close to the photosphere of the stars. The authorsclaimed that the formation of the first aromatic ring begins at 1.4 R∗ and, at aradius of 1.7 R∗ corresponding to a temperature of around 1700K, the conversionof single rings to PAHs begins. The resulting temperature is much higher thanthe temperature window calculated in previous studies. Cau (2002) investigatedthe contribution of periodic shocks to the formation of PAHs and their dimers inthe inner atmosphere of IRC+10216. He has shown that the amount of PAHs anddimers produced is not enough to explain the formation of carbon dust in the at-mosphere of the star, but it can account for the formation of the small disorderedcores of the kind observed in presolar grains. These cores might then grow largerthrough direct chemical reactions with acetylene. The yield of PAHs and dimersobtained was lower than that found by Frenklach & Feigelson (1989).

4.2 Soot formation in terrestrial combustion processes

Most experimental data on the kinetics of soot formation were obtained in flamesand shock waves. A rough picture of soot formation in flames has been provided byRichter & Howard (2000) who distinguish three main steps of the soot formation.The process starts with the formation of the molecular precursors of soot which


are supposed to be big PAHs of molecular weights larger than 300 u. The growthprocess from small molecules such as benzene to larger PAHs comprises both theaddition of C2, C3 or other small units to PAH radicals and reactions among thegrowing aromatic species.

The second part of the process is called particle inception. In this process,the conversion from molecular to particulate systems happens. The third part ofthe soot formation is the particle growth process which can be again sub-dividedinto the growth of particles by addition of gas-phase molecules, such as PAHs, andcoagulation via reactive particle-oparticle collisions, a process which significantlyincreases particles sizes. Whereas the first and third step is rather well understoodthe soot nucleation or particle inception is not yet understood. The first and thesecond step of soot formation are the most interesting ones for astrophysics.

However, one has to keep in mind that soot formation pathways in flames aremore complicated than in the more isothermal laser pyrolysis or laser ablationprocesses. Within a flame the temperature is rising from the bottom to the topand from the outer edge to the middle of the flame resulting in different formationpathways to soot particles. The application of gas-phase condensation methodswhich provides higher homogeneity in the temperature distribution are necessaryto obtain more general information on carbon grain formation in such processes.

In our laboratory, we have employed gas-phase condensation experiments suchas laser pyrolysis and laser ablation in quenching gas atmospheres to condensecarbonaceous particles within two different temperature regimes, a high- (HT)and a low-temperature (LT) condensation regime. More experimental details onthe techniques can be found in Jager et al. (2008) and (2009). In both cases, thegas-phase condensation is comparable to the grain formation processes suggestedto be relevant in astrophysical environments.

The study of the formation pathways has been performed by analytical char-acterization of the complete carbonaceous condensate, including high-resolutiontransmission electron microscopy (HRTEM), chromatographic methods, and massspectroscopy assuming that some fraction of the precursor molecules and interme-diate products are preserved in the final condensate due to an efficient quenching ofthe condensed carbon materials after the formation. In LT condensation processesat temperatures lower than 1700K, either mixtures of insoluble grains and PAHsor exclusively PAHs are formed. The typical internal structure of the particlescan be described by a concentrical arrangement of rather plane graphene layersaround a kind of disordered core (see Fig. 2). Matrix-assisted laser desorption andionization combined with time-of-flight (MALDI-TOF) mass spectrometer stud-ies have demonstrated that PAHs up to masses of 3000u and even higher areformed during the condensation process (see Fig. 3). All sizes of PAHs are formedin such condensations but no saturated hydrocarbon molecules could be identi-fied. LT condensation favors a formation process with PAHs as precursors andparticle-forming elements, which condense on the surface of carbonaceous seedsvia physi- and chemisorption. As a result, large carbonaceous grains are formedrevealing well developed planar or only slightly bent graphene layers in their in-terior. Since larger PAHs have a lower volatility compared to the smaller ones, a


Fig. 2. HRTEM micrograph of a grain formed in a LT (left) condensation process. The

average length of graphene layers is around 2 nm comparable to PAHs with masses of

around 1300 u (smaller PAH molecule). The largest layers have a size of 3 nm which

corresponds to PAHs of about 2700 u (large molecule). The HT condensed grains in the

right panel are built of fullerene fragments or bucky onions with nested fullerene cages.

Fig. 3. MALDI TOF spectrum of a LT soot sample showing PAHs with masses up to

3000 u. To exemplify a PAH molecule with comparable mass, a molecule comprising 91

condensed rings is presented.

preferred accumulation of large molecules during the surface growth process canbe observed. The accumulation of the large molecules on the surfaces of the seedscan be confirmed by a detailed analysis of the HRTEM micrographs (see Fig. 2)revealing a mean length of the graphene sheets inside the particles of around 2 nmcorresponding to PAHs with masses of around 1300u. The longest graphene layersdetermined in the soot particles have extensions of about 3 nm corresponding tothe highest-mass PAHs detected with MALDI-TOF measurements (Jager et al.2009).


The generation of fullerenes and fragments of them in the HT condensationprocess at temperatures higher than 3500K could be verified by using electronmicroscopy which can be shown in Figure 2. The presence of symmetric and elon-gated fullerene molecules with different numbers of carbon atoms is clearly visiblein the HRTEM micrographs (Jager et al. 2009). Fullerene fragments, fullerenesand buckyonions serve as building blocks for the very small carbon grains formedin such HT regimes. MALDI-TOF studies of the condensate have shown that noPAHs are formed in this process pointing to a completely different soot forma-tion pathway via polyyne chains, fullerenes and fullerene snatches. The formationpathway is found to be initiated by long and branched carbon macromolecules sug-gested to be the precursors for cyclic structures (Irle et al. 2003). The existence ofpolyyne structures in the condensate could be verified by in situ IR spectroscopythat shows the typical IR stretching bands of triply bound ≡CH and C≡C (Jageret al. 2008).

4.3 The nucleation process

The controversial part of the soot formation even in the combustion communityis the so-called inception, the formation of the first stable nuclei, i.e. the sootprecursor particles.

Most of the combustion models favor the influence of PAHs and the forma-tion of PAH clusters as soot nuclei (Dobbins et al. 1998; Richter & Howard 2000;Mansurov 1998). Chen & Dobbins (2000) considered large PAHs and PAH clusterswith layer structure as nucleation seeds. Schuetz & Frenklach (2000) discussed thedimerization of PAHs whereas Homann (1998) prefers the theory of PAH oligomer-ization resulting in the formation of so-called aromers. In PAH dominated sootformation processes, PAH radicals formed by hydrogen abstraction are discussedto be the driving force for the cluster formation (Richter et al. 2004). Reactions ofPAH radicals with PAHs and between PAH radicals were found to be the dominantpathway to soot nuclei. In flames, the transition from gaseous to solid particleswas assumed to depend on the temperature and pressure. Particles heavier than2000u and diameters of around 1.5 nm are considered as first solids (Lafleur et al.1996; Harris & Weiner 1988; McKinnon & Howard 1992).

The formation of larger, three-dimensional species with amorphous carbonstructure has been proposed by D’Alessio et al. (1998) and Minutolo et al. (1999).The role of fullerenes as soot nuclei (Zang et al. 1990) and the parallel growth offullerenes and large PAHs in flames (Grieco et al. 2000; Lafleur et al. 1996) havebeen favored by other groups.

Generally, the exact transition range from molecules to solid grains is differentfor different materials and varies between clusters of 100 and 1000 atoms. A niceexample for very small solid carbonaceous grains are the meteoritic nanodiamondswith a medium size of 1.3 nm corresponding to number of around 150 atoms. Thispoints to a high stability of these “clusters” consisting of sp3 hybridized carbonexclusively. Transitional objects between molecules and solids in HT condensateshave been observed by using HRTEM. Their structures are similar to the carbon


5 nm5 nm

1.7nm

5 nm5 nm

1.7nm

Fig. 4. Small transitional objects between molecules and solid soot grains observed in

HT condensation experiments.

cluster with spiral structures (Haberland 1994) which are supposed to be inter-mediates of the soot formation in condensation processes with high initial carbonatom densities. Three-dimensional fragments of fullerenes, oligomers of fullerenesor fragments and small nested buckyonions have been identified in this condensate(Jager et al. 2009).

5 Conclusion and outlook

Gas-phase condensation studies to uncover the formation process of carbon grainsat different condensation conditions are necessary to understand the structural andspectral properties of nanometer-sized carbon grains with respect to an applicationto cosmic carbon. So far, these studies succeeded in the production and structuralcharacterization of a big variety of carbon structures and in the tuning of specialstructural and spectral properties of the material.

A first basic understanding of the grain formation process in gas-phase con-densations was reached. Laboratory studies have shown that in flames as wellas in condensation processes with isothermal conditions, such as laser ablation orpyrolysis, multiple soot formation processes occur. In isothermal condensations, ahigh- and a low-temperature soot formation process have been identified in whicheither small fullerene-like carbon grains are produced via carbon chains, fullerenesand fullerene fragments, or soot grains with plane graphene layers are formedvia PAHs, respectively. However, we still need more insight into the complicatenucleation process of soot grains, in particular in PAH dominated formation pro-cesses. Further laboratory experiments to study the inception or nucleation stephave to be developed by coupling selective production methods with sophisticated


analytical tools (Biennier et al. 2010). In high-temperature condensation pro-cesses, three-dimensional fragments of fullerenes and small nested buckyonionshave been identified as transitional objects between molecules and solids.

References

Ajayan, P., Iijima, S., & Ichihashi, T., 1993, Phys. Rev. B, 47, 6859

Allain, T., Sedlmayr, E., & Leach, S., 1997, A&A, 323, 163

Banhart, F., 1999, Rep. Prog. Phys., 62, 1181

Biennier, L., Sabbah, H., Chandrasekaran, V., et al., 2011, this volume

Cau, P., 2002, A&A, 392, 203

Chen, H.X., & Dobbins, R., 2000, Combust. Sci. Tech., 159, 109

Cherchneff, I., Barker, J., & Tielens, A., 1992, ApJ, 401, 269

Cherchneff, I., & Cau, P., 1999, in ed. T. Le Betre, A. Lebre, and C. Waelkens, Asymp-totic Giant Branch Stars; Proceedings of the IAU Symposium 191 (San Francisco,ASP), 251

Cherchneff, I., 2011, this volume

Chhowalla, M., Wang, H., Sano, N., et al., 2003, Phys. Rev. Lett., 90, 155504

D’Alessio, A., D’Anna, A., Gambi, G., & Minutolo, P., 1998, J. Aerosol. Sci., 29, 397

Dartois, E., & Godard, M., 2011, this volume

Dobbins, R.A., Fletcher, R.A., & Chang, H.-C., 1998, Comb. Flame, 115, 285

Duley, W., 1994, ApJ, 430, L133

Elias, D.C., Nair, R.R., Mohiuddin, T.M.G.J., et al., 2009, Science, 323, 610

Frenklach, M., & Feigelson, E., 1989, ApJ, 341, 372

Gail, H.-P., & Sedlmayr, E., 1985, A&A, 148, 183

Geim, A.K., & Novoselov, K.S., 2007, Nature Materials, 6, 183

Grieco, W.J., Howard, J.B., Rainey, L.C., & Vander Sande, J.B., 2000, Carbon, 38, 597

Haberland, H., 1994, Clusters of Atoms and Molecules (Springer Verlag, Heidelberg)

Haddon, R.C., 1993, in ed. H. Kroto and D. Walton, The fullerenes (CambridgeUniversity Press, Cambridge)

Harris, P.J.F., 2009, Carbon Nanotube Science (Cambridge University Press)

Harris, S.J., & Weiner, A.M., 1988, Proc. Combust. Inst., 22, 333

Heidenreich, R., Hess, W., & Ban, L., 1968, J. Appl. Cryst., 1, 1


Hess, W., & Herd, C., 1993, in ed. J. Donnet, R. Bansal, and M. Wang, Carbon Black(Marcel Dekker, Inc., New York), 89

Homann, K.-H., 1998, Angew. Chem., 110, 2572

Iijima, S., 1991, Nature, 354, 56

Irle, S., Zheng, G., Elstner, M., & Morokuma, K., 2003, Nano Lett., 3, 1657

Jager, C., Henning, T., Spillecke, N., & Schlogl, R., 1999, J. Non-Cryst. Solids, 258, 161

Jager, C., Huisken, F., Mutschke, H., Llamas Jansa, I., & Henning, T., 2009, ApJ, 696,706

Jager, C., Kovacevic, E., Mutschke, H., Henning, T., & Huisken, F., 2010, in preparation


Jager, C., Mutschke, H., Henning, T., & Huisken, F., 2008, ApJ, 689, 249

Jones, A.P., Duley, W.W., & Williams, D.A., 1990, QJRAS, 31, 567

Kratschmer, W., Lamb, L.D., Fostiropoulos, K., & Huffman, D., 1990, Nature, 347, 354

Kroto, H., Heath, J., O’Brien, S., Curl, R., & Smalley, R., 1985, Nature, 318, 162

Kuc, A., Heine, T., & Seifert, G., 2010, Phys. Rev. B, 81, 085430

Lafleur, A.L., Taghizadeh, K., & Howard, J.B. e. a., 1996, J. Amer. Soc. Mass Spectrom.,7, 276

Lodders, K., & Fegley, B., J., 1999, in ed. T. Le Bertre, A. Lebre, and C.Waelkens, Asymptotic Giant Branch Stars; Proceedings of the IAU Symposium 191,SAO/NASA Astrophysics Data System, 279

Mansurov, Z.A., 1998, Combust Flame, 115, 285

McKinnon, J.T., & Howard, J.B., 1992, Proc. Combust. Inst., 24, 965

Minutolo, P., Gambi, G., D’Alessio, A., & Carlucci, S., 1999, Atmospheric Environment,33, 2725

Papoular, R., Breton, J., Gensterblum, G., et al., 1993, A&A, 270, L5

Pendleton, Y.J., & Allamandola, L.J., 2002, ApJS, 138, 75

Ravagnan, L., Bongiorno, G., Bandiera, D., et al., 2006, Carbon, 44, 1518

Ravagnan, L., Siviero, F., & Lenardi, C. e. a., 2002, Phys. Rev. Lett., 89, 285506

Richter, H., Granata, S., Green, W.H., & Howard, J.B., 2004, Proc. Combust. Inst., 30,1397

Richter, H., & Howard, J.B., 2000, Prog. Energy Combust. Sci., 26, 565

Robertson, J., 1986, Adv. Phys., 35, 317

Robertson, J., & O’Reilly, E., 1987, Phys. Rev. B, 35, 2946

Sakata, A., Wada, S., Okutsu, Y., Shintani, H., & Nakada, Y., 1993, Nature, 301, 493

Savchenko, A., 2009, Science, 323, 589

Schnaiter, M., Mutschke, H., Dorschner, J., Henning, T., & Salama, F., 1998, ApJ, 498,486

Schuetz, A.C., & Frenklach, M., 2000, Combust. Sci. Tech., 159, 109

Tauc, J., Grigorovici, R., & Vancu, A., 1966, Phys. Status Solid, 15, 627

Tomita, S., Fujii, M., & Hayashi, S., 2002, Phys. Rev. B, 66, 245424

Ugarte, D., 1992, Nature, 359, 707

Wada, S., Kaito, C., Kimura, S., Ono, H., & Tokunaga, A.T., 1999, A&A, 345, 259

Wu, J., Pisula, W., & Mullen, K., 2007, Chem. Rev., 107, 718

Zang, Q.L., O’Brien, S.C., & Heath, J.R. e. a., 1990, J. Phys. Chem., 90, 525


PAHS AND ASTROBIOLOGY

L.J. Allamandola1

Abstract. In dense molecular clouds, the birthplace of stars and plan-ets, interstellar atoms and molecules freeze onto extremely cold dustand ice particles. These ices are processed by ultraviolet light and cos-mic rays forming hundreds of far more complex species, some of astro-biological interest. Eventually, these rain down on primordial planetswhere they take part in the young chemistry on these new worlds.

Although the IR spectroscopy and energetic processing of interstel-lar ice analogs have been studied for nearly 30 years, similar studiesof PAH containing ices have only just begun. This paper presents re-cent results from laboratory studies on the vacuum ultraviolet (VUV)photochemistry of PAHs in water ice at low temperatures to assess theroles they play in the photochemical evolution of interstellar ices andtheir relevance to astrobiology. A number of “surprises” were found inthese studies on PAH containing water-rich ices, indicating that PAHslikely play very important, unexpected roles in cosmic ice chemistry,physics and astrobiology.

1 Introduction

The story of PAHs and astrobiology is an important part of a complete revolu-tion in our understanding of chemistry and biochemistry throughout the cosmos.Space was considered chemically barren for most of modern human history. Thatspell was broken some fifty years ago with the discovery of species such as NH3,H2CO, NH2CHO, and CO by radio astronomers during the late 1960’s and early1970’s (Cheung et al. 1968; Snyder et al. 1969; Rubin et al. 1971; Wilson et al.1970). Today there are almost 120 known interstellar species, the largest of whichis HC11N. Add to this the discovery of the Unidentified Infrared (UIR) EmissionBand spectra during the 1970’s, the suggestion that the UIR bands originated inunbelievably large interstellar PAHs during the 1980’s and the growing realization

1 NASA Ames Research Center, Space Science & Astrobiology Division, MS 245-6, MoffettField, CA 94035, USA



over the last twenty years that PAHs are widespread and very abundant through-out much of the Cosmos and it is clear that we no longer simply see the Universeas an H dominated physicist’s paradise.

While most of the work presented at this wonderful symposium has to do withusing the emission from vibrationally excited, gas phase PAHs as probes of differentenvironments, this paper discusses PAHs frozen in ices in dense molecular cloudsbecause it is in these clouds that stars, planets and moons form. These PAHs arepart of the extraterrestrial feedstock that contributes to the early chemistry andperhaps biochemistry on the primordial planets and moons.

In dense molecular clouds, atoms and molecules freeze onto the extremely colddust and ice particles forming mixed molecular ices. Infrared spectroscopy of theclouds reveals the composition of these ices (e.g. van Dishoeck 2004). With theoccasional exception of CO, the abundance of the identified species is much higherin the ice than in the gas phase. Water is by far the major ice component, withfrozen CO, CO2, and CH3OH often, but not always, present at concentrationsof roughly 10 to 20%. The ices also contain NH3, H2CO and CH4 at about thefew percent level. Within these clouds, and especially in the vicinity of star andplanet forming regions, these ices are processed by ultraviolet light and cosmicrays forming hundreds of far more complex species, some of biogenic interest.This process is illustrated in Figure 1. Unidentified IR spectral structure indicatesthat other species are also present at the few percent level (e.g. Gibb & Whittet2002; Boogert et al. 2008). Some of these will likely be produced via in-situ UVphotolysis of the interstellar ices. Although altering only a few percent of the ice,in-situ photochemistry favors chemical complexity and produces molecules andchemical groups within the ice that cannot be made via gas phase and gas-grainreactions. As these materials are the building blocks of comets and related tocarbonaceous micrometeorites, they are likely to be important sources of complexorganic materials delivered to habitable planets such as the primordial Earth. Thischapter focuses on the unique roles that PAHs play in the chemical evolution ofthese cosmic materials and their relevance to astrobiology.

2 Interstellar ice photochemistry and the production of biogenicmolecules

This section summarizes the photochemistry of mixed molecular and PAH contain-ing interstellar ice analogs comprised of the most abundant interstellar ice species,with emphasis placed on the production of biogenically interesting molecules.

There is a large body of work on UV driven photochemistry in various lowtemperature (10–30 K) ice mixtures without PAHs (e.g. Allamandola et al. 1988;Bernstein et al. 1995; Greenberg et al. 1995; Gerakines et al. 1996, 2001). As theseices are warmed, the starting materials and most volatile photoproducts sublime,leaving non-volatile residues on the substrate. Mass spectroscopic analysis showsthat hundreds of complex new species make up these residues (Greenberg et al.2000; Dworkin et al. 2004). Of the overwhelming array of compounds produced

L.J. Allamandola: PAHs and Astrobiology 307

Fig. 1. The Greenberg model of interstellar ice mantle formation and chemical evolu-

tion. The mantle grows by condensation of gas phase species onto the cold dust grains.

Simultaneously, surface reactions between these species, ultraviolet radiation and cosmic

ray bombardment drive a complex chemistry. Figure reproduced from Bernstein et al.

(1999a).

from even the simplest starting ice (H2O, CH3OH, NH3, and CO), only a fewhave been identified (Bernstein et al. 1995). In keeping with their expected lowconcentration and abundance with respect to the simpler ice species, clearcut IRspectroscopic evidence for these types of compounds in interstellar ices is presentlylacking although, as mentioned above, weak spectral structure in the 5–8 μm regionof protostars is consistent with their presence (e.g. Gibb & Whittet 2002; Boogertet al. 2008).

The residue is particularly important for astrobiology because it is quite plau-sible that this type of material is closely related to that preserved in comets,meteorites and interplanetary dust particles (IDPs), and it is believed that thesesources deliver between 12 and 30 tons of organic material to Earth monthly (Love& Brownlee 1993). During the period of great bombardment some 4 billion yearsago, the amount of extraterrestrial organic material brought to the early Earth wasmany orders of magnitude greater. Thus, this organic residue could have playedan important role in steering the early chemistry on habitable bodies such as theprimordial Earth.

Motivated by the possibility that the residues may contain compounds thatimpact early Earth biochemistry, experiments have been directed to searchingfor classes of biogenically important species. Remarkably, although all of thestarting compounds are water soluble, as shown in Figure 2, non-soluble dropletsspontaneously form when the residues are added to liquid water (Allamandolaet al. 1997; Dworkin et al. 2001). Droplet formation shows that some of thecomplex organic compounds produced by UV in these interstellar ice analogs are


Fig. 2. Water insoluble, fluorescent droplets produced from the residue when it is added

to water. The two images are of the same field.

amphiphilic, i.e. they have both a polar and non-polar component. These arethe types of molecules that make up cell membranes, and membrane productionis considered a critical step in the origin of life. These droplets are true vesicles(hollow droplets) with their interiors separated from the surroundings by theirlipid multilayer. Vesicle formation is thought critical to the origin of life sincevesicles provide an environment in which life can evolve, isolating and protectingthe process from the surrounding medium (Deamer et al. 2002).

Figure 2 also shows that the membranes trap photoluminescent molecules thatare also formed within the ice by UV irradiation (Allamandola et al. 1997; Dworkinet al. 2001). Thus, not only are vesicle forming compounds produced from thesimplest and most abundant interstellar starting materials, complex organics whichabsorb low energy UV are also made. The ability to form and trap energy receptorswithin these structures is considered another critical step in the origin of life as itprovides the means to harvest energy available outside the system. Interestingly,the vesicles formed from photoproducts of UV irradiated interstellar ice analogs areremarkably similar to those produced by extracts from the Murchison meteorite(Dworkin et al. 2001; Deamer et al. 2002).


Another class of potentially important organic molecules brought to the Earthby meteorites are amino acids, the molecular building blocks of proteins and en-zymes. The amino acids in meteorites are extraterrestrial as demonstrated bytheir deuterium enrichments – the highest of any species measured in a meteorite(Pizzarello & Huang 2005). Since laboratory photolysis experiments of plausiblepresolar ice mixtures also produce amino acids and closely related compounds(Bernstein et al. 2002; Munoz Caro et al. 2002; Agarwal et al. 1985), presolar icephotochemistry may be a source of these extraterrestrial amino acids.

As with other polyatomic interstellar species, PAHs should also become partof the ices in dense clouds. Weak, broad absorption bands in molecular cloudspectra are consistent with PAHs frozen in ice mantles at the few percent level,however they suffer severe blending with the strong absorption bands of the majorice components (Keane et al. 2001; Bernstein et al. 2005; Sandford et al. 2004;Bouwman et al. 2011; Bouwman et al., elsewhere in this volume). The chemistry ofPAHs in cosmic ice analogs has also been investigated to understand the roles PAHsmight play in astrochemistry and astrobiology. Photolysis of interstellar ice analogscontaining pure PAHs does not destroy the parent PAH, but decorates it withchemical side-groups formed from the photoproducts of the other ice constituents.For example, in water ice, oxidized products found in the residues include aromaticethers, alcohols, and ketones with structures as shown in Figure 3 (Bernsteinet al. 1999b). Hydrogen atom addition transforms some of the edge rings intocyclic aliphatic hydrocarbons, creating molecules with both aromatic and aliphaticcharacter and decreasing the overall degree of aromaticity. Likewise, UV andproton irradiation of PAHs in NH3, CH4, CH3OH, HCN and CO2 ices producesaromatic structures decorated with amino (-NH2), alkyl (-CH3, -CH2), hydroxyl(-OH), cyano (-CN), and carboxy (>CO) side groups, again species similar to thecompounds found in meteorites (Bernstein et al. 2003, 2002, 2001). Most recently,photolysis of the nitrogenated single ring aromatic molecule, pyrimidine, in waterice produced the nucleobase, uracil (Nuevo et al. 2009). These changes open upan aromatic based biogenic chemistry that is not available to the parent PAH orin non-PAH containing ices. For example, compounds closely related to aromaticketones are widely used in current living systems for electron transport across cellmembranes (hardware) while nucleic acids provide the backbone for informationcoding in DNA and RNA (software).

From both the astrochemical and astrobiological perspective, the productionof such an interesting array of aromatic based compounds warranted a deeperunderstanding of the processes that occur within the ice during the UV irradi-ation of PAHs in cosmic ice analogs. However, because of severe line blending,the “traditional” laboratory approach using IR spectroscopy to follow chemicalchange within irradiated ices cannot provide this information (Bernstein et al.2005; Sandford et al. 2004; Bouwman et al. 2011; Bouwman et al., elsewhere inthis volume). It is for this reason that optical studies have been undertaken tofollow the stepwise reaction pathways and kinetics of PAHs in water ice and to de-termine important properties such as ionization efficiencies, onset temperature of


Fig. 3. PAH structures produced by UV irradiation of PAH:H2O ices at interstellar

temperatures.

reaction during warm-up, possible identities of some of the reaction intermediates,and so on. The next section summarizes this work.

3 Photochemistry and photophysics of PAHs in cosmic ice analogs

A number of “surprises” were found in these studies on the photochemical pro-cessing of PAH containing water-rich ices. Foremost among these is that rapidphotoionization is the dominant process that occurs when dilute, low temperaturePAH/H2O ices are initially exposed to UV radiation (Gudipati & Allamandola2003; Bouwman et al. 2010). As illustrated in Figure 4, a large fraction (generally50–70%) of the parent PAH is converted to PAH+ within minutes of UV onset. Ifradiation is stopped at this point, the PAH cation appears to be stable indefinitely(on laboratory timescales) as long as the ice temperature is not raised. The PAHcation is also stabilized by the H2O ice to remarkably high temperatures. Thisis shown in Figure 5 which compares the spectrum of the quaterrylene (QTR,C40H20) cation in water ice at 20 K to that of the same sample at 120 K aftera nearly one month, very slow, stepwise warm-up. The QTR+ band only startsto weaken after the ice matrix is warmed above 120 K, but the loss is slow up to


Fig. 4. Spectral evolution of H2O/4-methylpyrene (4MP, 1500/1) ice at 15 K as a func-

tion of VUV irradiation time showing the loss of neutral 4MP and simultaneous growth

of 4MP+ with VUV exposure. The dot-dashed line shows the absorbance spectrum be-

fore photolysis. The black, red, green, and blue solid lines show the loss of the neutral

4MP band and the simultaneous growth of 4MP+ after 30, 300, 1200, and 4800 s, re-

spectively, of in-situ VUV (vacuum UV) irradiation. The vibronic bands peaking near

340 nm correspond to neutral 4MP, and those near 450 nm to the 4MP+ photoproduct

(Figure reproduced from Gudipati & Allamandola 2003; see also Bouwman et al. 2010).

∼140 K. Although the loss increases at higher temperatures, band reduction stopswhen warm-up stops. Remarkably, the band is still evident at 170 K, the temper-ature at which the ice sublimes (Gudipati & Allamandola 2006). In addition, theionization potential (IP) of PAHs trapped in water ice at low temperatures are atleast 2 eV lower than the corresponding gas phase IPs (Gudipati & Allamandola2004; Woon & Park 2004)

These initial studies showed that PAHs likely play important but overlookedroles in cosmic ices, especially regarding the chemistry of charged species. Afar more detailed understanding of the processes involved is necessary before onecan incorporate these ideas into astrochemical models. To provide this criticalinformation, Bouwman et al. (2009, 2010) and Bouwman et al. in press, haveundertaken a systematic study of several PAHs in H2O ice, tracking the in-situbehavior of each PAH and their photoproducts within the ice itself as a function


Fig. 5. The 520 to 940 nm (0.52 to 0.94 μm) spectrum of quaterrylene (QTR, C40H20, up-

per right insert) in a water ice matrix. The spectrum of the freshly deposited QTR/H2O

(∼1/500) ice at 20 K is shown in the upper left insert. The full spectrum (solid line)

is that of the same ice after photolysis with 266 nm laser light. Negative bands rep-

resent depletion of the parent neutral QTR and positive bands represent generation of

QTR+. The dotted-line spectrum was recorded at 120 K, after step-wise warm-up over

28 days. The QTR+ band areas are equal, showing QTR+ is stable in water-ice at this

temperature (Gudipati & Allamandola 2006).

of ice temperature and UV irradiation time with sub-second responsivity. Figure 6shows the dramatic effect ice temperature has on the photochemistry of the PAHpyrene (Py, C16H12) in water ice and Figure 2 in Bouwman et al., elsewhere in thisvolume, shows the evolution of each band as a function of photolysis time. Theability to measure spectra simultaneously with photolysis and with sub-secondtime resolution permits kinetic studies previously inaccessible and provides newinsights into the processes occurring within the ice during photolysis (Bouwmanet al. 2009, 2010; Bouwman et al. in press). Since these temperatures span therange of dust temperatures in dense clouds in the transition zones with star andplanet forming regions and the ranges in the temperatures of icy Solar Systembodies beyond the orbit of Jupiter, PAH driven ice photochemistry is likely toimpact a variety of cosmic environments.


Fig. 6. The VUV induced spectroscopic changes in Pyrene:H2O ice at 25 (bottom)

and 100 K (top) as a function of photolysis time. Temperature plays a critical role

in determining photochemical pathways for PAH containing ice. The loss of the neutral

pyrene (Py, C16H12) band complex near 330 nm is similar at both temperatures. However,

in the 25 K ice, cation formation (∼450 nm) is favored over the range of species responsible

for the broad, 350 to 500 nm absorption and the 400 and 405 nm band carriers. The

opposite holds for the 100 K ice. Figure from Bouwman et al. (2010).

4 Astrobiology

The similarities between interstellar ice analog photoproducts with the extraterres-trial materials delivered to the Earth strengthens the case for taking interstellarPAHs and ices into account when pondering the exogenous contribution to theorigin of life that is illustrated in Figure 7.


Fig. 7. The cycle of interstellar material that leads to the production and evolution

of complex organic compounds that can be delivered intact to, and mixed with, those

produced on planets and moons. There they participate in the primordial chemistry

on these young objects. Figure courtesy Jason Dworkin, NASA Goddard Space Flight

Center, adapted from Deamer et al. (2002).

Three possible roles that these interstellar materials might have played in theorigin of life on Earth are:

(1) Supplier of basic, prebiotic raw materials from which biotic compoundswere eventually produced on Earth. Comets, IDPs and meteorites were very likelyresponsible for raining a rich inventory of prebiotic molecules down onto the earlyEarth. Recall the variety of species that make up the bulk of interstellar ices.


(2) Source of complex prebiotic materials poised to play a direct role once in afavorable environment. The work described here supports the idea that meteorites,dust, and comets could deliver species which are sufficiently complex to play adirect role in the chemistry of the origin of life. From this perspective, cometsmight represent quite potent prebiotic cocktails. All the complex species discussedabove are produced by the irradiation of simple ice mixtures that are widespreadthroughout dense molecular clouds. Thus, it seems reasonable to hypothesizethat wherever a planet or satellite is formed in the habitable zone, it will accretethese chemical remnants from the nascent cloud from infalling IDPs, cometary(formerly interstellar) and meteoritic materials. One of the challenges before us isto understand this chemistry and to assess the relative importance of this inputinto the origin of life versus an origin from the compounds produced on, or alreadypresent on, the primordial planet.

(3) Fountainhead of species undergoing the basic processes of life. Going onestep further, this article ends on a speculative note. The results summarized hereare quite startling when viewed from the perspective held not too long ago thatchemistry in space was very limited. We have just started to scratch the surface ofthis complex chemistry, we know there are hundreds of residue compounds not yetanalyzed and we have just realized that ion-moderated processes also play a role inice chemistry. Given our level of ignorance at this stage, an intriguing possibilityis the production, within the processed mixed molecular ices themselves – whetherin the interstellar medium or a comet – of species poised to take part in the lifeprocess, or perhaps even at the earliest stages of what could be perceived of asrudimentary reactions of a living system.

This “jump-starting” of the life process by the introduction of such marginallybiologically active species into the “warm pond” on a habitable planet may not beas far-fetched as it would at first seem. Although one could hardly imagine a betterdeep freeze than isolation within a comet, with temperatures measured in tens ofdegrees K, there are repeated episodes of warming for periodic comets. Since thesewarming episodes, which can be repeated many times in a given comet, can lastanywhere from periods of weeks to several months, there is ample time for a veryrich mixture of complex organics to develop even though these warming episodesare sandwiched between long periods of extreme cold. At this early stage in ourunderstanding of cometary ices it already appears that comets contain many ofthe types of compounds which are considered important players in the origin oflife, compounds which play central physical roles and compounds which play keychemical roles. Indeed there is evidence for nearly every class of organic chemicalthought important in the “RNA world” in these irradiated ices and they are all inclose proximity to one another. We are just beginning to appreciate the chemicalcomplexity and possibilities of these ices. Thus, it is no longer inconceivable, atleast to this author, that comets played a far more active role in the origin oflife than simply that of a spectator-delivery system of raw materials. PerhapsDarwin’s “warm little pond” is a warmed comet.

Summing up, in cold molecular clouds, the birthplace of planets and stars, in-terstellar molecules freeze into ice particles comprised of water, methanol, ammonia,


carbon monoxide, carbon dioxide and PAHs. In these clouds, especially close tostar and planet forming regions, these ices are processed by UV light, cosmic rays,radical surface reactions, etc. forming far more complex species, many of biogenicinterest. During star and planet formation, many of these compounds becomeincorporated in comets and meteorites which eventually seed primordial planetswhere they take part in the budding chemistry on these young worlds.

While I have been blessed to have worked with many outstanding experimentalists and theo-reticians over the years, I particularly wish to thank three exceptionally gifted experimentalistswho made this chapter possible: Murthy Gudipati, Jordy Bouwman, and Harold Linnartz. Theyshined the first light into the microscopic details of PAH/ice processing. I am especially gratefulto Harold Linnartz and NASA for making it possible for me to spend six wonderful months work-ing in Leiden once again and Christiaan Boersma who, under severe time pressure, helped LaTeXthis article. I also acknowledge support from NASA’s Long Term Space Astrophysics, Exobiology,and Astrobiology programs for their continued support throughout this field’s blossoming.

References

Agarwal, V.K., Schutte, W., Greenberg, J.M., Ferris, J.P., et al., 1985, Orig. Life, 16, 21

Allamandola, L.J., Sandford, S.A., & Valero, G.J., 1988, Icarus, 76, 225

Allamandola, L.J., Bernstein, M.P., & Sandlord, S.A., 1997, IAU Colloq. 161: Astron.Bioch. Orig. Search Life Universe, 23

Bernstein, M.P., Sandford, S.A., Allamandola, L.J., Chang, S., & Scharberg, M.A., 1995,ApJ, 454, 327

Bernstein, M.P., Sandford, S.A., & Allamandola, L.J., 1999a, Scient. Amer., 281, 26

Bernstein, M.P., Sandford, S.A., Allamandola, L.J., et al., 1999b, Science, 283, 1135

Bernstein, M.P., Dworkin, J.P., Sandford, S.A., & Allamandola, L.J., 2001, Meteor.Planet. Sci., 36, 351

Bernstein, M.P., Dworkin, J.P., Sandford, S.A., Cooper, G.W., & Allamandola, L.J.,2002, Nature, 416, 401

Bernstein, M.P., Elsila, J.E., Dworkin, J.P., et al., 2002, ApJ, 576, 1115

Bernstein, M.P., Moore, M.H., Elsila, J.E., et al., 2003, ApJ, 582, L25

Bernstein, M.P., Sandford, S.A., & Allamandola, L.J., 2005, ApJS, 161, 53

Boogert, A.C.A., et al., 2008, ApJ, 678, 985

Bouwman, J., Paardekooper, D.M., Cuppen, H.M., Linnartz, H., & Allamandola, L.J.,2009, ApJ, 700, 56

Bouwman, J., Cuppen, H.M., Bakker, A., Allamandola, L.J., & Linnartz, H., 2010, A&A,511, A33

Bouwman, J., Mattioda, A.L., Linnartz, H., & Allamandola, L.J., 2010, A&A, 511, A33

Cheung, A.C., Rank, D.M., Townes, C.H., Thornton, D.D., & Welch, W.J., 1968, Phys.Rev. Lett., 21, 1701

Deamer, D., Dworkin, J.P., Sandford, S.A., Bernstein, M.P., & Allamandola, L.J., 2002,Astrobiology, 2, 371

Dworkin, J.P., Deamer, D.W., Sandford, S.A., & Allamandola, L.J., 2001, Proc. Nat.Acad. Sci., 98, 815

Dworkin, J.P., Seb Gillette, J., Bernstein, M.P., et al., 2004, Adv. Space Res., 33, 67


Gerakines, P.A., Schutte, W.A., & Ehrenfreund, P., 1996, A&A, 312, 289

Gerakines, P.A., Moore, M.H., & Hudson, R.L., 2001, J. Geophys. Res., 106, 33381

Gibb, E.L., & Whittet, D.C.B., 2002, ApJ, 566, L113

Greenberg, J.M., Li, A., Mendoza-Gomez, C.X., et al., 1995, ApJ, 455, L177

Greenberg, J.M., et al., 2000, ApJ, 531, L71



Gudipati, M.S., & Allamandola, L.J., 2006, ApJ, 638, 286

Keane, J.V., Tielens, A.G.G.M., Boogert, A.C.A., Schutte, W.A., & Whittet, D.C.B.,2001, A&A, 376, 254

Love, S.G., & Brownlee, D.E., 1993, Science, 262, 550

Munoz Caro, G.M., et al., 2002, Nature, 416, 403

Nuevo, M., Milam, S.N., Sandford, S.A., Elsila, J.E., & Dworkin, J.P., 2009, Astrobiology,9, 683

Pizzarello, S., & Huang, Y., 2005, Geochim. Cosmochim. Acta, 69, 599

Rubin, R.H., Swenson, G.W., Jr., Benson, R.C., Tigelaar, H.L., & Flygare, W.H., 1971,ApJ, 169, L39

Sandford, S.A., Bernstein, M.P., & Allamandola, L.J., 2004, ApJ, 607, 346

Snyder, L.E., Buhl, D., Zuckerman, B., & Palmer, P., 1969, Phys. Rev. Lett., 22, 679

van Dishoeck, E.F., 2004, ARA&A, 42, 119

Wilson, R.W., Jefferts, K.B., & Penzias, A.A., 1970, ApJ, 161, L43

Woon, D.E., & Park, J.-Y., 2004, ApJ, 607, 342


SOLID STATE MOLECULAR REACTORS IN SPACE

M. Maurette1

Abstract. Lunar minerals and impact glasses, convert the polyatomicbeam of solar wind (SW) ions into a flux of small molecules (e.g., H2,N2, H2O, CO, CO2, CH4, C2H4, C2H6, HCN, metal carbides and deu-terides, etc.). They thus behave as “Solid State Molecular Reactors”.Moreover, ∼100–200 μm size micrometeoroids (μMs) have also beenexposed to the SW in the zodiacal cloud, before being captured by theEarth and recovered as Antarctic micrometeorites. They are mostlycomposed of a PAH-rich hydrous-carbonaceous material, which ampli-fies their power as molecular reactors. In particular, during the first∼200 Myr of the post-lunar period, about 75% of the μMs have beenmelted and/or volatilized upon atmospheric entry. The release of theirvolatile species triggered a cosmic volcanism around the mesopausethat ruled the formation of the early Earth’s atmosphere and climate.Furthermore, a fraction of the μMs that survive unmelted upon at-mospheric entry did settle on the proto-oceans floors. Upon furtherburial in sediments their constituent PAH-rich kerogen was crackedinto abiotic oil, which generated giant oil slicks that fed prebiotic chem-istry. Many stars, of all ages and types, are embedded into a secondarydebris-disk loaded with ion implanted μMs. Some of them are expelledto the interstellar medium (ISM) where they behave first as “dormant-invisible” molecular reactors, until they became reactivated by variousprocesses to synthesize interstellar molecules. This short paper onlyfocus on some highlights of this research dealing with the synthesis ofimportant interstellar molecules, including the most abundant ones (H2

and CO) and H2O, HCN and PAHs, all involved in prebiotic chemistry.

1 Introduction: Organic geochemistry of lunar molecular reactors

Lunar dust grains are unique reference samples for learning how to extrapolatethe results of laboratory simulation experiments to cosmic dust grains, exposedin ultra-high vacuum (∼10−13 torr) to beams of slow and fast ions. Left panel of

1 CSNSM, University Paris XI, Batiment 104, 91405 Orsay-Campus, France



Figure 1 shows the dark-field electron micrograph of a μm-size crystalline lunarfeldspar grain, which was observed with a 3 MV high voltage transmission electronmicroscope by Christian Jouret. Radiation damage features due to the solar wind(SW) and solar energetic particles (SEPs) can be observed, which include -(i)-a darker rim with an average thickness of ∼40 nm that decorates an amorphouscoating (AC) of SW radiation damage, and -(ii)- a high density of latent tracks ofSEPs iron-group nuclei that are best observed as faint lines of dark contrast withinthe dotted yellow ellipse. The right panel of the same figure shows a zoomed–updetail of another grain, with the contrast enhanced to make the ion tracks morevisible.

The identification of these features requires artificial ion implantations of ter-restrial lunar analogs with both SW type ions (1 keV/amu) and fast Fe ionsE∼1MeV/amu). Upon in-situ thermal annealing in the HVEM, the fossil SW-ACsand SEPs-tracks, as well as the artificial SW-ACs and Fe-tracks, did anneal at thesame temperature (∼700–800 ◦C). Moreover, the average thickness of the fossilACs, of about 40 nm (see Figs. 1a, 1d, in Maurette 1976), corresponds to boththat observed during artificial 1 keV/amu SW-implants, and to the projected dam-age range, Rpd ∼ 40 nm, of 1 keV/amu SW ions in silicates, as estimated withthe SRIM software.

In 1969, organic geochemists discovered that lunar dust grains returned by theApollo 11 mission did only release small molecules (H2O, CO, CH4, and HCN,etc.) during either pyrolysis at moderate temperature (600 ◦C) or HF etching.They suggested that these molecules were likely produced during the implantationof high fluences of SW ions in the grains. We next helped testing the validityof this assumption while exposing successively ∼1 cm2 flat surfaces of terrestrialanalogs (ilmenite, oligoclase and a simulated lunar glass) to high fluence implants(∼1015 to 2×1018 cm−2) of “marked” isotopes of hydrogen (2D), carbon (13C)and nitrogen (15N) ions, at energy of ∼ 1 keV/amu. The group of Al Burlingame(Space Science Laboratory, UC Berkeley) next showed (see Bibring et al. 1974)that the mixture of small molecules released upon heating and/or HF etching fromthese targets are very similar to the natural blend released from lunar dust grains(Simoneit et al. 1973).

2 Desorption of molecules during ion implants

In 1975, Plasma Desorption Mass Spectrometry (see section 4, last paragraph)was already used to show that ∼ 100 MeV fission fragments from 252Cf directlydesorb DOH from D-implanted muscovite mica (Maurette et al. 1975). To fur-ther investigate this stimulated desorption, another set of the pre-implanted lunaranalogs have been exposed to the 10 keV primary beam of Ar of the ionic analyzerbuilt by G. Slodzian. This instrument gives the depth concentration profiles of allspecies released from the targets. J.P. Bibring observed a large variety of synthe-sized species ranging from 13C-Mg to 13C3D8, and including D2, 15N2, D2O, and13CO. However, 15NO and 15NO2 were not detected (15NO/13CD ≤ 10−4).

M. Maurette: Solid State Molecular Reactors in Space 321

Fig. 1. Left: lunar Micron-size molecular reactor. The shard of impact glass stuck to

the main grain has not been noticeably eroded by SW sputtering. Thus it has still not

reach the critical saturation SW fluence to function as a molecular reactor. Right: detail

of another dust grain with enhanced contrast to make ion tracks more visible to the

untrained eye (the thin dark lines in the image).

This study could be extended to a primary beam of 3 keV deuterium, throughthe decisive help of W. Moller and his team (Max-Plank-Institute fur Plasma-physik, Garching). The three lunar analogs, as well as olivine, silicon, sapphireand a soda-lime glass, have been run with a sophisticated reaction chamber de-veloped for understanding the interactions of a hot DT fusion plasma (T = 3He)with the first wall of fusion reactors, at D and T energy in the keV to MeV range.This chamber is equipped with a quadrupole mass spectrometer on-line with botha 30 kV ion accelerator for deuterium implantation, and a 2.5 MV Van de Graaffaccelerator for the analysis of the implanted targets with a 750 keV 3He beam,through the D(3He, α)H nuclear reaction. This reaction yields the residual con-centration and depth concentration profile of the implanted deuterium. The massspectrometer was tuned to detect simultaneously the release of D2 and D2O, whichare considered as stumbling blocks in either the gas phase chemistry of the ISM(H2) or prebiotic chemistry (H2O).

This on-line study reveals the intimate functioning of the seven molecular reac-tors (Borgesen et al. 1986). In a first group, including ilmenite, olivine, the alkali-poor LSG (0.02% K2O), sapphire and silicon, a complete retention of D is observedas soon as the deuterium beam is turned on around a fluence of ∼ 1015 cm−2.At increasing fluence, a sudden onset of deuterium re-emission (as D2) is foundaround 1 × 1017 cm−2, which finally increases to 80–100% of the implanted deu-terium around a saturation fluence, Φs ∼ 1018 cm−2. These major characteristicsof the deuterium release have also been previously reported for metals (at low


temperature), and other targets at room temperature, including semiconductors,fused silica and a sintered powder of graphite, which did release both D2, CD, C2Dand CD4. During all the re-emission stage in the O-rich targets, the incoming Dyields a D2O / D production ratio of 10−4, 3×10−4 and 2×10−5, for ilmenite,oligoclase and the SLG, respectively.

However, both the Na-rich oligoclase and soda-lime glass show a very distinctbehavior, as they start loosing (i.e., synthesizing) D2 at the onset of implantation(1015 cm−2). They would be the best molecular reactors in a low ion flux environ-ment. On a flat target, perpendicular to the anisotropic SW beam at 1 AU, thisfluence is reached in a few months. However, a spherical grain on top of the surfaceof the Moon would require a longer exposure to accumulate the same fluence oneach element of its surface, because the grain is partially shielded from the SW,e.g., during the lunar night.

3 Hydrous-carbonaceous molecular reactors on the early earth

In 1975, we already suspected that lunar type molecular reactors might be effec-tive in the ISM (cf., last section in Maurette & Price 1975). In 1987, a muchmore powerful family of ∼100–200 μm size molecular reactors (i.e., the hydrous-carbonaceous μMs) was discovered as Antarctica micrometeorites (see Maurette2009, for the most recent review about collects and applications). Their markedsimilarities with the Wild-2 dust particles returned by the Stardust mission inJanuary 2006 (Dobrica et al. 2009), further indicate that they likely originatefrom comets. This is further supported by IR observations, which revealed thatcontrarily to previous thoughts, comets are very efficient “dust machine gun” inthe solar system (Lisse et al. 2004). Observations of comets P/Encke and P/IRASwith IR telescopes on the ISO and MSX spacecrafts show dust-to-gas mass lossratios about 50 times larger than those inferred from measurement in the visible.These high mass-loss ratios, which were reconfirmed for comet Temple 1 duringthe Deep Impact mission, clearly show that comets are the major contributors ofdust in the interplanetary medium.

After ejection, most of these 100 μm size cometary dust grains have been tem-porarily stored in the Zodiacal Cloud for duration of ∼200 000 yr (Raisbeck &Yiou 1989), and thus implanted with SW- and SEPs-ions. This cloud is probablya very faint remnant of the secondary debris-disk of the early Sun, which did ex-ponentially decay until about 4 Ga ago, while triggering a period of early heavybombardment of the Earth-Moon system (Greaves 2005).

About 99% of the Antarctica micrometeorites are only related to the mostvolatile-rich hydrous-carbonaceous chondrites, which are relatively rare (about2–3% of the meteorite falls). Their major carbonaceous component is a kerogencomposed of a mixture of PAH moieties and aliphatic hydrocarbons, and their con-stituent water (10 wt.%) is mostly stored in their major hydrous silicate, saponite,that starts dehydrating at ∼100◦C. Upon atmospheric entry μMs would thus re-lease solar wind implanted neon, but also molecules originating from their kerogen(N2 and CO2), hydrous silicates (H2O), and metallic sulfides (SO2).


These volatiles ended up making the early Earth’s atmosphere (i.e., beforethe formation of O2), during the first ∼200 Myr of the post-lunar period, afterthe giant impact that did form the Moon (Maurette 2009). This impact did alsoclose the formation time interval of the Earth, and blew-off the intractable pre-lunar atmosphere of the young Earth at a time when the Earth’s mantle hasalready been extensively degassed. Therefore, after the impact, there was notenough volatile species left in the mantle to feed the formation of the massiveearly Earth’s atmosphere, which required the delivery of total amounts of H2O,CO2 and SO2 corresponding to partial pressures of about 270, 60 and 80 bars,respectively (today, CO2 and SO2 are mostly stored as carbonates in the crustand metallic sulfides in the upper mantle, respectively). This μMs origin of theatmosphere is strongly supported by the Ne/N2 and the D/H ratios of Antarcticmicrometeorites. They best fit the corresponding ratios observed in the air and theoceans, respectively, which are incompatible with all other models of atmosphereformation.

Moreover, a fraction of the μMs that did survive unmelted upon atmosphericentry with their full load of PAH-rich kerogen landed on the proto-oceans floors.They well fit the definition of good “source rocks” for the “economic exploita-tion” of petroleum, i.e., fine clays with at least 0.5 wt.% of kerogen (Rondeel2002). Therefore, they likely generated giant abiotic oil slicks that assisted prebi-otic chemistry (Dias & Maurette 2007). In 1971, Dave Holland and collaborators(Lasaga et al. 1971) proposed a distinct type of primordial oil slick. But theiroil was produced during the solar UV irradiation of a primordial methane andnitrogen atmosphere, which had soon to be disregarded.

4 Kerogen-rich molecular reactors in the ISM

Frank Shu (1982) pointed out that the extinction of visible starlight could becaused by “bricks”. However, he cautioned that bricks cannot produce a reddeningof starlight. The combination of true absorption and reddening of this starlight isthus attributed to grains that are slightly smaller (≤0.4 μm) than the wavelengthsof visible light in order to scatter blue light more efficiently than red light. Thisgeneral extinction does not exclude the presence of ≥100 μm size “bricklets”, whichwere injected from various stellar sources in the ISM. They could be made of thelarge diversity of solids that have the potential of desorbing molecules upon ionimplants (Sect. 2), including hydrous carbonaceous micrometeoroids (see Sect. 3).All of them would start as somewhat dormant and invisible molecular reactors inthe ISM, until they are re-activated by various processes (see next section).

In 1994, 15 Antarctica micrometeorites (i.e., μMs) have been ranked on ascale of increasing thermal metamorphism upon atmospheric entry (Clemett et al.1996). The PAH moieties of their constituent kerogen have been desorbed withthe IR laser pulses of a double laser microscope equipped with a reflectron time-of-flight mass spectrometer (μL2MS). A great diversity of PAH mass spectra wereobserved, from the simplest spectra of unmelted μMs (similar to that of the CM2-type Murchison chondrite) to the complex spectra of partially melted μMs that


mimic that of soot particles (see Figs. 4a to 4d, in Clemett et al. 1996). “Exotic”PAHs were also observed, such as vinyl PAHs that might have promoted interestingorganometallic reactions.

McFarlane and co-workers invented the 252Cf plasma desorption mass spec-trometer, which allowed the determination of the molecular weight of large organicmolecules, such as insulin, which desorbs “intact” from thin deposits irradiatedwith ∼100 MeV 252Cf-fission fragments (Torgeson et al. 1974). Therefore, fastions should also softly desorb PAHs from kerogen, like the IR laser pulses of theμL2MS. We commented (Maurette et al. 2003) on the in-situ FTIR analysis offive organics irradiated with 10 MeV/amu 40Ar ions, under oxygen free conditions(Reynaud et al. 2001). These materials loose the memory of their initial proper-ties when the radiation dose exceeds ∼300 Mega-Grays, i.e., after an exposure ofabout 104 yr to SEPs at 1 AU. They are all converted into an insoluble materialdubbed as kerogen, which might desorb its PAH moieties in the ISM.

5 Summary and prospects

In the last 70 yr, an abundant literature was published about the main ingredi-ents of dust evolution models in galaxies, including the synthesis of interstellarmolecules (cf., Dwek et al. 2009). Our studies show that many reaction channelsare opened upon implantation of a polyatomic beam of fast ions in a great di-versity of Solid State Molecular Reactors. A challenge is to check whether theireasy-to-desorb and abundant species (e.g., H2, CO and PAHs), can help simplify-ing the chemical models proposed for the synthesis of interstellar molecules in thegas-phase, “which contain thousands of reactions, many of which have not beenstudied theoretically or in the laboratory” (quote from E.Herbst).

Another challenge is to identify the dominant process that can activate molec-ular desorption from the “dormant” molecular reactors in the ISM. All previousprocesses, including those listed by Maurette & Price in 1975, might be fully super-seded by the high flux of low energy (2–20 MeV/amu) galactic cosmic rays (LOWs)if McCall and co-workers are right (McCall 2010). They reported challenging pa-pers (e.g., Indriolo et al. 2009) dealing in particular with the determination of thehigh degree of ionization of the diffuse ISM (as inferred from the H+

3 ion) and theproduction rates of Li, Be, B and gamma-rays. They tested multiple LOWs energyspectra and found only two spectra that best fitted the corresponding observations.

I selected their simplest “broken power-law” distribution of cosmic rays in rel-ativistic momentum, reproduced in Figure 2, that represents a differential numberflux, Φp [particles cm−2s−1sr−1 (GeV/n)−1]. Nick Indriolo has kindly computedthe corresponding integrated flux of protons, in the three energy ranges consid-ered in this paper: Δ1 (0.002–1GeV/n); Δ2 (2–20 MeV/n) and Δ3 (5–20 MeV/n).The corresponding “integral” proton fluxes (number cm−2 s−1 sr−1 ) are about3.7, 1.5, and 0.9, for Δ1, Δ2, and Δ3, respectively. For grains smaller than theprojected range of 2 MeV/n LOWs in silicates (about 42 μm and 18 μm for H and


Fig. 2. Possible energy spectrum of low energy cosmic rays in the ISM (Courtesy N. In-

driolo).

Fe, respectively), and which include all “visible” grains with sizes ≤1 μm in theISM, this value has to be multiplied by 4π because the LOWs flux is isotropic.

The lifetime of a large dust grain in the ISM is about 3×107 yr, cycling severaltimes between dense and diffuse clouds (see e.g. Tielens 2005). With the resultingflux of 20 protons cm−2s−1(in the 2–20 MeV/n range) the kerogen simulationexperiments discussed in section 4 predict that the organic component of the visiblegrains would be transformed into kerogen in ∼105 yr of exposure in the diffuseinterstellar clouds. Furthermore, artificial implants of Fe and Ar ions, show thatcrystalline silicates, would be rendered amorphous by a fluence of ≥1011 VH cm−2

at the very low temperatures of the ISM, which quench the diffusion inducedannealing of atomic defects. Considering a low energy cutoff value of 5 MeV/nrequired for the penetration of VH ions in the diffuse clouds, the Indriolo predictedproton flux (11 cm−2 s−1) corresponds to an exposure of ∼3×107 yr and ∼106 yrin the ISM, for a solar VH/H ratio (10−5) and the 30 times higher ratio measuredfor ≥1 GeV/n GCRs, respectively. A reasonable VH/H ratio just a few timeshigher than solar would insure that in the diffuse ISM, crystalline silicates wouldbe rare but PAHs quite abundant.

The Mcfarlane desorption mass spectrometer can measure PAH desorptionyields of kerogen-rich materials, which are required to extrapolate the desorptionmodel to the ISM. The presence of abundant “bricklets” along the orbits of cometsis brightly signaled by meteor showers. However, the ISM bricklets have still tobe detected.


I am indebted to: IN2P3 for constant support; A. Burlingame, S. Clemett, C. Hammer, C. Jouret,G. Kurat, W. Moller, N. Indriolo and M. Pourchet for important contributions; E. Dobrica andC. Engrand for their characterization of Antarctic micrometeorites, and; C. Joblin for stimulatingthoughts, encouragements and help with LaTeX formatting.

References

Bibring, J.P., Burlingame, A., Chaumont, J., et al., 1974, Geochim. Cosmochim. Acta,2, 1747

Borgesen, P., Maurette, M., Moller, W., & Monart, B., 1986, in Radiation effects inInsulators 3, ed. I.H. Wilson & R.P. Webb, Gordon and Breach Science Publishers),711

Clemett, S.J., Chillier, X.D.F., Gillette, S., et al., 1996, Origins life Evol. Biosphere, 28,425

Dias, F., & Maurette, M., 2007, Lunar Planet. Sci., 37, A1268

Dobrica, E., Engrand, C., Quirico, E., Montagnac, G., & Duprat, J., 2009, MAPS, 44,1643

Dwek, E.E., Galliano, F., & Jones, A., 2009, ASP Conf. Ser., 414, 183

Greaves, J.S., 2005, Science, 307, 68

Indriolo, N., Fields, B.D., & McCall, B.J., 2009, ApJ, 694, 257

Lasaga, A.C., Holland, H.D., & Dwye, M.J., 1971, Science, 174, 53

Lisse C.M., VanCleve, J., Adams, A.C., et al., 2004, Icarus, 171, 444

Maurette, M., & Price P.B., 1975, Science, 187, 121

Maurette, M., 1976, Nucl. Inst. Meth., 132, 579

Maurette, M, Banifatemi, A., Della-Negra, S., & Lebeyec, Y., 1983, Nature, 303, 159

Maurette, M., Balanzat, E., & Duprat J., 2003, Lunar Planet. Sci., 34, A1743

Maurette M., 2009, ASP Conf. Ser., 414, 137

McCall, B.J., 2010, in http://bjm.scs.uiuc.edu/

Raisbeck, G.M., & Yiou, F., 1989, Meteoritics, 24, A318

Reynaud, J., et al., 2001, Spectrochim. Acta, A57, 797

Rondeel, H.D., 2002, in www.geol.vu.nl

Shu F.H., 1982, The Physical Universe: An Introduction to Astronomy (UniversityScience Books), 1

Simoneit, B.R., Wszolek, P.C., Christiansen, P., Jackson, R.F., & Burlingame, A.L.,1973, Geochim. Cosmochim. Acta, 37, 1603

Tielens, A.G.G.M., 2005, “The Physics and Chemistry of the Interstellar Medium”(Cambridge University Press), Section 13

Torgeson D.F., Skowronsky, R.P., & Macfarlane, R.D., 1974, Biochem. Biophys. Res.Commun., 60, 616


POLYCYCLIC AROMATIC HYDROCARBONSAND THE EXTINCTION CURVE

G. Mulas1, G. Malloci1, C. Joblin2, 3 and C. Cecchi–Pestellini1

Abstract. Aromatic carbon, in some form, has been an essential in-gredient by and large in all models of the extinction curve, since theoriginal proposal to attribute the bump at 217.5 nm to “astronomicalgraphite”. This aromatic carbon is most naturally identified, in up todate models, with a population of Polycyclic Aromatic Hydrocarbons(PAHs), free and/or clustered. In all models, this PAH population ac-counts for the far–UV nonlinear rise in the extinction curve, contributesto the bump and possibly part of the large set of unidentified, discreteabsorption features in the visible (the Diffuse Interstellar Bands). Wereview the current state of our understanding of the contribution ofPAHs to interstellar extinction, and what constraints can be imposedon the PAH population by fitting extinction models to observations.

1 Interstellar extinction before PAHs

Interstellar extintion has a long, winding story. It dates as far back in time as1774, with sir William Herschel noticing a region in the Scorpio constellationremarkably devoid of stars, which he called a “hole in the sky”. The fact thatsuch regions were a common occurrence in the Milky Way was clear in the earlyyears of 1900, with the systematic observations of Barnard, as was stated in atextbook of the time (Clerke 1903), and it began to be hypothesised that theywere not actually holes in the fabric of the sky, but instead might be “obscured”by some intervening material. For some decades, this hypothesis was the topic ofsome heated debates, some of which are on record in the proceedings of academicmeetings of the time (Shapley & Curtis 1921). It was only about in 1930 that firmevidence of interstellar extinction became available (Trumpler 1930).

1 Istituto Nazionale di Astrofisica – Osservatorio Astronomico di Cagliari – strada 54, localitaPoggio dei Pini, 09012 – Capoterra (CA)2 Universite de Toulouse, UPS, CESR, 9 Av. colonel Roche, 31028 Toulouse Cedex 4, France3 CNRS, UMR 5187, 31028 Toulouse, France



Thus far, interstellar extinction was accessible to study mainly in the visi-ble part of the spectrum, where it was known to increase linearly for decreasingwavelength, when expressed in magnitudes vs. wavenumbers (hence the name“reddening” still widely used for it), and was therefore believed to be essentiallydue to scattering and absorption by some form of classical dust particles. Theprecise composition and structure of this astronomical dust was not really con-strained, as about any distribution of classical particles made of any refractorymaterial would produce a linear extinction curve in the visible. It was only aboutin the 1940 decade that van de Hulst produced the first full fledged model of in-terstellar extinction based on well–defined dust particles (van de Hulst 1949). Theonly structure known at that time in this almost linear behaviour of extinctionwas a set of optically thin, unidentified features, called the Diffuse InterstellarBands (DIBs). The first ones were discovered by Merrill (1934), their number hasbeen growing steadily with the increasing sensitivity of instruments and currentlyexceeds four hundreds (Hobbs et al. 2010), still with no unambiguous identifi-cation despite some tentative ones. Since DIBs are the subject of some othercontributions (Snow & Desheed, Cox in this volume), we will only briefly mentionthem here. For a beautiful and thorough review of the early studies of interstellarextinction, see e.g. Greenberg & Shen (1999).

The opening of the UV section of the spectrum to astronomical observations,with the advent of rocket– (Stecher 1965) and space–based (see e.g. Bless & Savage1970, and references therein) instruments, revealed more structure in the thus farrelatively simple extinction curve. The slope of its linear behaviour changed inthe near UV, and on top of it appeared two strong features: a broad Drude–likeabsorption “bump” peaking at about 217 nm, followed by a non–linear additionalrise in the extinction at still lower wavelengths in the far–UV. The phenomeno-logical behaviour of the extinction curve in the visible and UV was summarised inthe monumental, systematic works by Fitzpatrick & Massa (1986,1988,1990,2005,2007,2009). They collected the observed extinction curves for several hundreds oflines of sight, and found that in the UV (i.e. for λ < 270 nm) they could all beparametrised satisfactorily with a relatively simple functional form.

k(λ − V ) =E(λ − V )E(B − V )

={

c1 + c2x + c3D (x, x0, γ) x ≤ c5

c1 + c2x + c3D (x, x0, γ) + c4(x − c5)2 x ≥ c5(1)

where x = λ−1 and

D (x, x0, γ) =x2

(x2 − x20)

2 + x2γ2.

This functional form represents the “normalised” color excess k(λ−V ), consideredas a function of inverse wavelength, with the superposition of a linear contribution,a Drude profile and a quadratic term. At longer wavelengths, k(λ − V ) is verynearly linear through the visible range, with a change of slope at around 440 nm,and tends to zero for very long wavelengths. The extinction “bump”, which is thesingle most intense interstellar absorption feature, was found to have an almostinvariable peak position, whereas its intensity and full width at half–maximum

G. Mulas et al.: PAHs and the Extinction Curve 329

Fig. 1. Average interstellar extinction curve in the galactic plane, as given by Fitzpatrick

& Massa (2007). The definition of k(λ−V ) is given in Equation (1). The weak structure

in the visible range is the set of Diffuse Interstellar Bands, taken from Hobbs et al. (2010),

and arbitrarily enhanced by a large factor to make them visible on the scale of this plot.

vary substantially from one line of sight to another. The average extinction curvein the galactic plane is shown in Figure 1. The overall shape of the extinctioncurve in the UV suggested the presence of a large population of very small dustparticles, the so called “Platt particles” after their original proposer (Platt 1955).It was then suggested, soon after the availability of the first rocket data, thatsuch particles could actually consist of some sort of aromatic carbon (Donn 1968).Indeed, from the integrated intensity of the bump it is straightforward to obtaina robust constraint on the abundance of its carrier (see e.g. Draine 1989)

nx

nHfx = 9.3 × 10−6,

wherenx

nHis the abundance of the “bump” carrier, and fx the oscillator strength of

the transition producing it. It is apparent that even assuming a very large oscillatorstrength fx only the most abundant elements can possibly produce the bump. Ifone takes fx ∼ 0.16, appropriate for intensity per C atom of the very strong π∗ ← πresonance in graphitic materials, producing the bump observed in the averagegalactic extinction curve requires about 1

4 of the available interstellar carbon. Noother carriers could realistically produce such a strong feature compatibly withinterstellar abundances.


The above argument, together with the resemblance of the observed “bump”profile with the π∗ ← π bands in graphite, led to the development of dust modelsattempting to fit the entire interstellar extinction curve (Mathis et al. 1977; Lee &Draine 1984; Draine 1987a). These models assumed power–law size distributionsof spherical silicate and graphite particles, and used Mie theory, together with“optimised” optical constants (Lee & Draine 1984; Draine 1985; Draine 1987b),obtaining a very successful fit to the observed extinction curves. One drawbackwas that the “optimised” optical constants did not match any precise known ma-terial, but were obtained by collecting the (then available) experimental data, withsubtle ad hoc changes to make sure that they joined smoothly, that they were con-sistent with the Kramers–Kronig relations, and that they fitted the astronomicalobservations.

2 PAHs enter the scene

The opening of another previously unavailable spectral window, again with space–based observations, was key to the appearance of PAHs in astronomy. The firstIR observations showed a large excess of emission in the 12 μm IRAS band, withrespect to the others at longer wavelengths. IR spectra, furthermore, showed anubiquitous, relatively invariant spectrum of bright emission bands at ∼3.3, 6.2,7.7, 8.6, and 11.3 μm. Sellgren (1984) was the first to notice that especially the3.3 μm band could not be accounted for by equilibrium thermal emission of dustparticles, since it would require unrealistically high temperatures. She hinted,remembering the proposed role of temperature fluctuations in the evaporation ofice mantles of dust grains (Purcell 1976; Greenberg & Hong 1974), that these bandsmight be due to a population of particles with such a small thermal capacity tobe transiently heated to the required high temperatures by a single UV photon.Sellgren’s proposal, in hindsight, was reminiscent of the early suggestion by Platt(1955) and Donn (1968) for extinction (“Platt particles”),

Soon thereafter, two groups independently proposed that these “particles”could be free–flying Polycyclic Aromatic Hydrocarbons (PAHs) (Leger & Puget1984; Allamandola et al. 1985), since they appeared to fit all the required char-acteristics: they absorb effectively UV radiation, convert it with high efficiencyto vibrational excitation, and release it as IR emission, in bands qualitativelymatching the observed “Unidentified Infrared Bands” (UIRs). They are also veryresistant to photodissociation (provided they are large enough), and can possiblyform in C–rich outflows of evolved stars, with chemical pathways inspired fromcombustion chemistry. Shortly thereafter, it was also realised that since interstel-lar PAHs, depending on size and charge state, would have also discrete bands inthe visible, this made them very plausible candidate carriers for DIBs (van derZwet & Allamandola 1985; Leger & d’Hendecourt 1985).

Purely energetic considerations, based on the overall integrated emission inthe UIRs, assuming unit conversion efficiency and UV absorption cross–sections(Puget & Leger 1989; Joblin et al. 1992), led to estimate that about 15–20% of


the available interstellar carbon must be locked in PAHs to produce the observedIR emission fluxes. This led to the consequence that PAHs must be taken intoaccount as a necessary ingredient of interstellar extinction models (Puget & Leger1989).

Indeed, from this point on, all interstellar extinction models got updatedto include some sort of “astronomical PAH” component (Desert et al. 1990;Siebenmorgen & Kruegel 1992; Li & Greenberg 1997; Dwek et al. 1997). In par-ticular, all models use these “astronomical PAHs” to account for the non–linearfar–UV rise, very effectively decoupling it from the extinction “bump” at 217 nm.At the time, PAHs were assumed to contribute only to the non–linear far–UVrise, while the bump was still entirely attributed to a population of classical small“astronomical graphite” particles. This was very convenient, as it provided a nicephysical explanation of the observed behaviour of these two components of theextinction curve, which were known to vary independently in observations.

The assumption that “astronomical PAHs” did not contribute to the bump wasdue to the following reasons. The bump position is known to be very nearly invari-ant in all lines of sight, whereas its intensity and width vary. It is also remarkablysmooth and no fine structure was detected on it so far. The (few) experimentalspectra of PAHs available, almost entirely neutral ones (Joblin et al. 1992; Joblin1992) showed π∗ ← π transitions with different positions, and with considerablestructure on a wavelength scale much smaller than the bump, even when mix-tures were considered, making PAHs an unattractive candidate to account for the217 nm extinction feature (Draine & Malhotra 1993; Mathis 1994).

Moreover, PAHs were believed to be largely ionised (see e.g. D’Hendecourt &Leger 1987; Verstraete et al. 1990; Salama et al. 1995) in the diffuse interstellarmedium. In their seminal model, Desert et al. (1990) assumed a mixture ofionised PAHs to have no absorption in the bump region, probably reasoning thatthe variation of the position of π∗ ← π transitions in different molecules wouldsmooth them out over such a wide wavelength range to make them irrelevant.

Direct measurements of the cross–sections of PAH cations were (and still are)rather complicated, since PAH cations are very reactive species. The first experi-ments used the technique of matrix isolation, in which a reactive species is includedas a trace component in a solid which perturbs it as little as possible (e.g. a frozennoble gas like argon or neon). Irradiation of the neutrals by far-UV photons leadsto a fraction of ionised species in the matrix.

Experiments like these produced a large amount of good data on the visiblespectra of PAH cations (e.g. Salama & Allamandola 1992a, 1992b), mainly aim-ing at the comparison with astronomical spectra of the DIBs. For this spectralregion the technique is rather effective, since the neutral PAHs studied have nofeatures in the visible whereas the cations do. Measuring the strong π∗ ← πtransitions in the bump region, however, is much more difficult, since the neutralprecursors (also present in the matrix) do have strong and broad transitions there,difficult to disentangle from the features of the cations. This probably misled Lee& Wdowiak (1993,1994a,1994b), who interpreted their experiments on irradiated


PAH–containing glasses as showing that the π∗ ← π transitions were substantiallysuppressed in cations.

For all these reasons, all interstellar extinction models, for the contribution ofPAHs to UV extinction, stuck with the assumption of Desert et al. (1990) andadopted an average cross–section of the neutral PAHs from which the π∗ ← πtransitions were simply removed, leaving only the non–linear rise in the far UVdue to the onset of σ∗ ← σ transitions (see e.g. Desert et al. 1990; Siebenmorgen& Kruegel 1992; Li & Greenberg 1997; Dwek et al. 1997).

Several years later, Duley & Seahra (1998) used the discrete dipole approx-imation (DDA) to study the spectra of PAHs, both single and in clusters, as afunction of hydrogenation and ionisation. They found that π∗ ← π transitionsappeared to be almost unaffected by ionisation, in stark contrast with the previ-ous interpretation by Lee & Wdowiak (1993,1994a,1994b) and the assumptions ofthe astronomical PAH models of the time. They also found that the peak of theπ∗ ← π transitions shifts systematically with dehydrogenation, and claimed thatheavily dehydrogenated PAHs fit the observed position of the interstellar extinc-tion bump. One year later, Chillier et al. (1999) measured the VUV spectrumof the perylene cation with an indirect technique, unambiguously confirming thatthe strong π∗ ← π transitions are almost unaffected by ionisation. In hindsight,this is hardly surprising, since, at least for large PAHs, one more or less electrondoes not overly perturb the π and π∗ orbitals.

This awareness that all PAHs have strong π∗ ← π transitions, and the real-isation that they do not leave much room in the interstellar carbon budget fora separate population of carbonaceous particles producing the bump, caused ageneral update of the photo–absorption cross–sections giving the contribution ofPAHs to interstellar extinction models (see e.g. Weingartner & Draine 2001; Li& Draine 2001; Draine & Li 2001; Draine in this volume). The far–UV part ofthe extinction curve, including the bump and the non–linear rise, were assumedto be completely insensitive to charge, and to include a strong π∗ ← π feature.As a consequence, models lost the flexibility to easily account for independentvariation of the bump and the far–UV non–linear rise, since from now on theywere assumed to be essentially due to the same population of molecules and tobe largely insensitive to size and charge of the PAHs. On the other hand, theonset of absorption by PAHs became dependent on charge, as populations of PAHcations were found to begin absorbing at lower energies with respect to neutrals,reflecting experimental and theoretical results (e.g. Ruiterkamp et al. 2002,2005).PAHs became by and large the main carrier of the 217 nm extinction bump (seee.g. Draine in this volume).

Since all PAHs were now believed to strongly absorb in the UV, Clayton et al.(2003) searched for weak structure in the UV extinction curve, using Hubble SpaceTelescope observations at high spectral resolution, in the hope of being able toidentify some features due to individual PAHs. This resulted in no detection, andin the placement of relatively low upper limits, meaning that no single PAH isvery abundant in the diffuse ISM.


As experimental techniques for gas–phase measurements, e.g. in cold jets,became more feasible, more and more experimental measurements of PAH spectrabecame available (see e.g. Brechignac & Pino 1999; Pino et al. 1999; Brechignacet al. 2001; Biennier et al. 2003; Biennier 2004; Tan et al. 2005; Staicu et al. 2006;Rouille et al. 2007; Salama et al. 2008), but mainly in the visible and near–UV,to compare them with observations of the DIBs. PAH spectra in the vacuum UV,especially for cations, being much more difficult and less promising for a directidentification, were left behind. A thorough review of the state of the art of thespectroscopy of PAHs is given in this volume by Pino et al..

3 Systematic theoretical spectra come in

In the same years, the steady increase in available computing power brought largerand larger PAHs within reach of quantum chemistry calculations, in particularusing the Time–dependent Density Functional Theory (TD–DFT). This techniquewas proven to hit a sweet spot between computational cost and accuracy, andbegan to be routinely used both as a complement to experimental measurements,to enable the interpretation of photo–absorption spectra and the identification ofbands, and to perform systematic studies on trends of the spectral behaviour offamilies of PAHs as a function of structure, size, ionisation state etc. (Ruiterkampet al. 2005). However, commonly available implementations of TD–DFT areincreasingly ineffective and/or inaccurate for transitions above the ionisation limitof a given molecule. This appears to be not the case for codes which implementTD–DFT in real time and represent wavefunctions by their values on a finite gridin real space, such as the open–source Octopus (Marques et al. 2003).

Malloci (2003) and Malloci et al. (2004) tested this code against the fewexperimental photo–absorption spectra available up to the vacuum UV, findingthat the accuracy of TD–DFT calculations remained by and large the same upto high energies (a few tenths of an eV for band positions) as for low energies.Moreover, the computational cost, using this kind of implementation, does notchange with the spectral range on which the spectrum is computed, making itparticularly suitable for computing whole PAH spectra up to (and even beyond)the Lyman limit.

Having validated the approach (cf. Fig. 2), Malloci and co–workers (Mallociet al. 2005, 2007a, 2007b proceeded to compute photo–absorption spectra of alarge number of PAHs over the whole wavelength range relevant for interstellarextinction, and made their results publicly available on a web–based database1.This made possible a systematic study of the effect of hydrogenation and chargestate. It was confirmed (Malloci et al. 2008) that dehydrogenation progressivelyshifts the π∗ ← π transitions bluewards, by more than 2 eV upon complete dehy-drogenation, confirming the trend previously predicted by Duley & Seahra (1998)and by Duley & Lazarev (2004). However, the increased accuracy of TD–DFT

1http://astrochemistry.oa-cagliari.inaf.it/database


Fig. 2. Average of the photo–absorption cross–sections computed via Time–Dependent

Density Functional Theory for a sample of neutral PAHs, normalised to the number of

C atoms (1 Mb = 10−18 cm2 = 10−2 A2). The theoretical curve, in continuous line, is

compared with the experimental spectrum (small crosses) of a mixture of small gas-phase

PAHs obtained by evaporation of a coal–pitch extract (Joblin et al. 1992).

calculations with respect to the DDA technique used by Duley & Seahra (1998)showed that if PAHs were severely dehydrogenated their π∗ ← π absorption wouldfall between the “bump” and the onset of the non–linear far–UV rise, placing astrong upper limit on the overall dehydrogenation of PAHs.

As to ionisation, it was demonstrated (Cecchi–Pestellini et al. 2008) that thereis indeed a relevant systematic effect: going from anions to neutrals to cationsand dications it appears that while the π∗ ← π transitions (in the “bump” region)are almost unchanged, the onset of the large, broad absorption due to σ∗ ← σtransitions is progressively shifted to higher energies, significantly reducing theabsorption in the wavelength range of the non–linear far–UV rise (see Fig. 3).This is important for interstellar extinction models, since it offers a physicallysensible parameter which can change the observed independent variation of the“bump” and non–linear far–UV rise.

4 Including actual PAH cross–sections in extinction models?

The availability of a relatively large number of real computed photo–absorptionspectra of PAHs made it possible to attempt, for the first time, the construction ofan interstellar extinction model including a distribution of classical dust particles


Fig. 3. Systematic variation of the average of the photo–absorption cross–sections of a

sample of computed PAHs (1 Mb = 10−18 cm2 = 10−2 A2), normalised to the number

of C atoms, with charge state (anions in light gray, neutrals in medium gray, cations in

dark gray, dications in black).

and a linear combination of cross–sections of a sample of PAHs in several chargestates (Cecchi–Pestellini et al. 2008). This has the obvious advantage of beinga more physical approach, and the obvious drawback of introducing a ludicrousnumber of free parameters (the column density of each species included in thelineal combination).

Comparison with observed extinction curves showed quantitatively severalpoints. The first is that it is indeed practically possible to match the extinc-tion curve very accurately using a mixture of even a relatively small sample of“real” PAHs. Some examples are shown in Figure 4. A few tens of species provideenough chemical diversity to very effectively wash out spectral features of individ-ual species in the UV, resulting in a very smooth extinction curve which matchesvery accurately the observed one in the UV, including both the “bump” and thefar–UV rise. While this would now appear obvious in hindsight, Figures 2 and 3make it evident that not every mixture of PAH spectra results in washing outspectral structure: an actual fit on real observations was necessary to prove thatobtaining a smooth bump with PAHs is not only possible but actually easy (the fitis very strongly underdetermined, see below). On the other hand, at least in somepeculiar lines of sight there must be some (probably carbonaceous) component ofvery small (nanometer sized) particles other than PAHs, as PAHs cannot repro-duce extreme cases with a strong non–linear far–UV rise and no bump. Severalworks showed that one feasible component with the required properties would be


Fig. 4. Comparison between some observed extinction curves (dotted lines, Fitzpatrick

& Massa 2007), spanning a large range of values of the total to selective extinction

ratio RV = AVEB−V

, and the fits (solid continuous line) obtained by Cecchi–Pestellini

et al. (2008). The dashed line shows the contribution of classical dust particles, the

dash–dotted lines the contribution of PAHs.

a carbonaceous material in which carbon has an sp3 structure, such as nanodia-monds or hyperhydrogenated PAHs (see e.g. Rai & Rastogi 2010).

The other important point is that it was quantitatively shown that even if a fitof the extinction curve with a mixture of PAHs is severely underconstrained, mean-ing that there is no uniquely defined composition of PAHs required to produce agiven observed interstellar extinction curve, nonetheless some collective propertiesof the mixture are well–determined, e.g. the charge per carbon atom, which relatesto the relative intensity between the “bump” and the non–linear far–UV rise. This


Fig. 5. Correlation between the charge (in units of electron charges) per C atom in PAHs

and the c4/c3 (c4 and c3 are defined in Eq. (1)) ratio. From top to bottom, this is shown

for PAH anions, neutrals, cations, dications, and the whole population. A clear trend is

visible on the bottom panel.

is illustrated in Figure 5, in which the lowest panel shows that increasing ratiosof the c4 (proportional to the intensity of the non–linear far–UV rise) versus c3

(proportional to the intensity of the bump) parameters in the Fitzpatrick & Massaparametrisation (cf. Eq. (1)) correspond to an increasing value of the charge percarbon atom in PAHs. The bottom panel in Figure 5 shows quantitatively thatwhen in an extincion curve the non–linear far–UV rise is more intense with respectto the bump, fitting it requires a mixture of PAHs with less electrons per C atom.

While such a “brute force” approach to the contribution of PAHs to interstellarextinction is probably not per se the best way for a comprehensive modellingof dust absorption and emission, it shows that the best currently available suchglobal model (e.g. Li & Draine 2001) lose some important information with their(necessary) simplification of PAH absorption, for example disregarding the impactof charge on far–UV extinction. As Einstein liked to say, a model should bemade as simple as possible without losing significant physics, but not simpler thanthat. Some more work still needs to be done to achieve this result for PAHs andextinction, to find an acceptable middle ground somewhere between the nightmareof a poorly defined mixture of too many PAHs and an unrealistic single spectrumof the “average astronomical PAHs”.


References

Allamandola, L.J., Tielens, A.G.G.M., & Barker, J.R., 1985, ApJ, 290, 25

Biennier, L., 2004, Chem. Phys. Lett., 387, 287

Biennier, L., Salama, F., Allamandola, L.J., & Scherer, J.J., 2003, J. Chem. Phys., 118,7863

Bless, R.C., & Savage, B.D., 1970, IAU Symp., 36, 28

Brechignac, P., Pino, T., & Boudin, N., 2001, Spectrochim. Acta, 57, 745

Brechignac, P., & Pino, T., 1999, A&A, 343, L49

Cecchi-Pestellini, C., Malloci, G., Mulas, G., Joblin, C., & Williams, D.A., 2008, A&A,486, L25

Chillier, X.D.F., Stone, B.M., Salama, F., & Allamandola, L.J., 1999, J. Chem. Phys.,111, 449

Clayton, G.C., et al., 2003, ApJ, 592, 947

Clerke, A.M., 1903, Problems in astrophysics (London, A. & C. Black)

Desert, F.-X., Boulanger, & F., Puget, J.L., 1990, A&A, 237, 215

D’Hendecourt, L.B., & Leger, A., 1987, A&A, 180, L9

Donn, B., 1968, ApJ, 152, 129

Draine, B.T., & Li, A., 2001, ApJ, 551, 807

Draine, B.T., & Malhotra, S., 1993, ApJ, 414, 632

Draine, B.T., 1989, Interstellar dust, Proc. IAU Symp., 135, ed. L.J. Allamandola &A.G.G.M. Tielens, 313

Draine, B.T., 1987, BAAS, 19, 1077

Draine, B.T., 1987, ApJS, 64, 505

Draine, B.T., 1985, ApJS, 57, 587

Draine, B.T., & Lee, H.M., 1984, ApJ, 285, 89

Duley, W.W., & Lazarev, S., 2004, ApJ, 612, L33

Duley, W.W., & Seahra, S., 1998, ApJ, 507, 874

Dwek, E., et al., 1997, ApJ, 475, 565

Ehrenfreund, P., D’Hendecourt, L., Joblin, C., & Leger, A., 1992, A&A, 266, 429

Fitzpatrick, E.L., & Massa, D., 2009, ApJ, 699, 1209


Fitzpatrick, E.L., & Massa, D., 2005, AJ, 130, 1127

Fitzpatrick, E.L., & Massa, D., 1990, ApJS, 72, 163



Greenberg, J.M., & Hong, S.-S., 1973, BAAS, 5, 380

Greenberg, J.M., & Hong, S.-S., 1974, Galactic Radio Astron., 60, 155

Greenberg, J.M., & Shen, C., 1999, Ap&SS, 269, 33

Hobbs., L.M., York, D.G., Thorburn, J.A., et al., 2010, A New Atlas of the DiffuseInterstellar Bands: HD 183143, 65th International Symposium On MolecularSpectroscopy



Joblin, C., 1992, Ph.D. Thesis

Lee, H.M., & Draine, B.T., 1984, BAAS, 16, 446

Lee, W., & Wdowiak, T.J., 1994, Molecules and Grains in Space, 312, 675

Lee, W., & Wdowiak, T.J., 1994, The First Symposium on the Infrared Cirrus andDiffuse Interstellar Clouds, 58, 295

Lee, W., & Wdowiak, T.J., 1993, ApJ, 410, L127

Leger, A., Verstraete, L., D’Hendecourt, L., et al., 1989, Interstellar Dust, 135, 173


Leger, A., & Puget, J.–L., 1984, A&A, 137, 5

Li, A., & Draine, B.T., 2001, ApJ, 554, 778

Li, A., & Greenberg, J.M., 1997, A&A, 323, 566


Malloci, G., 2003, Ph.D. Thesis



Malloci, G., Mulas, G., Cappellini, G., Fiorentini, V., & Porceddu, I., 2005, A&A, 432,585

Malloci, G., Mulas, G., Cecchi-Pestellini, C., & Joblin, C., 2008, A&A, 489, 1183

Marques, M.A.L., Castro, A., Bertsch, G.F., & Rubio, A., 2003, Computer Phys.Commun., 151, 60

Mathis, J.S., 1994, ApJ, 422, 176

Mathis, J.S., Rumpl, W., & Nordsieck, K.H., 1977, ApJ, 217, 425

Merrill, P.W., 1934, PASP, 46, 206

Pino, T., Boudin, N., & Brechignac, P., 1999, J. Chem. Phys., 111, 7337

Platt, J.R., 1955, ApJ, 123, 486

Puget, J.L., & Leger, A., 1989, ARA&A, 27, 161

Purcell, E.M., 1976, ApJ, 206, 685

Rai, R.K., & Rastogi, S., 2010, MNRAS, 401, 2722

Rouille, G., Arold, M., Staicu, A., et al., 2007, J. Chem. Phys., 126, 174311

Ruiterkamp, R., Cox, N.L.J., Spaans, M., et al., 2005, A&A, 432, 515

Ruiterkamp, R., Halasinski, T., Salama, F., et al., 2002, A&A, 390, 1153

Salama, F., Tan, X., Biennier, L., et al., 2008, BAAS, 40, 187

Salama, F., & Allamandola, L.J., 1992, ApJ, 395, 301

Salama, F., & Allamandola, L.J., 1992, Nature, 358, 42

Salama, F., Joblin, C., & Allamandola, L.J., 1995, P&SS, 43, 1165

Sellgren, K., 1984, ApJ, 277, 623

Shapley, H., & Curtis, H.D., 1921, Bull. Nat. Res. Council, 2, 217

Siebenmorgen, R., & Kruegel, E., 1992, A&A, 259, 614

Staicu, A., Krasnokutski, S., Rouille, G., Henning, T., & Huisken, F., 2006, J. Mol.Struct., 786, 105


Stecher, T.P., 1965, ApJ, 142, 1683

Steglich, M., Jager, C., Rouille, G., et al., 2010, ApJ, 712, L16

Tan, X., Biennier, L., Cami, J., & Salama, F., 2005, IAU Symp., 235, 70P

Trumpler, R.J., 1930, Lick Obs. Bull. XIV, 154

van de Hulst, H.C., 1957, Light scattering by small particles (Wiley)

van der Zwet, G.P., & Allamandola, L.J., 1985, A&A, 146, 76


Weingartner, J.C., & Draine, B.T., 2001, ApJ, 548, 296


THE DIFFUSE INTERSTELLAR BANDS IN HISTORYAND IN THE UV

T.P. Snow and J.D. Destree1

Abstract. The diffuse interstellar bands (DIBs) have been known ofsince 1922, but their carrier molecules remain unidentified to this day.We present a brief history of DIB observations, followed by a list ofconstraints any suggested origin must face, and finally a preview ofcurrent research for ultraviolet DIBs using the Cosmic Origins Spec-trograph on the Hubble Space Telescope. We conclude that PAHs areconsistent with all the listed constraints, but that PAHs may not bethe only molecular species responsible for the DIBs.

1 Introduction

The discovery of the first known diffuse interstellar bands (DIBs) was publishedin 1922 by Mary Lea Heger, then a graduate student at the Lick Observatory(Heger 1922). After almost a century, the DIBs are now recognized as a mainreservoir of organic material in the galaxy – but their origin remains unknown.Some characterize the DIBs as the oldest spectroscopic mystery in all of science.Now, PAHs are among the leading candidate as their carrier molecules.

Recent review papers are Herbig (1995), Snow (2002), Krelowski (2002), andSarre (2006). An on-line DIB bibliography, listing all published papers on DIBs,can be found at: http://dibdata.org/.

This paper is divided into sections as follows: a brief history of DIB obser-vations; the molecular origin of the DIBs; properties of the DIBs; the possiblerole of PAHs as carriers of the DIBs; and a report of the first search for UV DIBsdone with the Cosmic Origins Spectrograph (COS) recently installed on the HubbleSpace Telescope (HST).

1 Center for Astrophysics and Space Astronomy, Department of Astrophysical and PlanetarySciences, University of Colorado at Boulder, Campus Box 389, Boulder, CO 80309-0389, USA



2 History of DIB observations

In 1921, while Mary Lea Heger was making observations of “stationary” (i.e. in-terstellar) atomic sodium absorption lines in stellar spectra, she noted the presenceof two broad features at 5780 and 5797 A that appeared to be interstellar in origin(Heger 1922). Heger took little notice of these features, other than to tabulatethem, and it was more than a decade later when Paul Merrill began the first ofmany systematic studies of the bands.

At about the time when Merrill became actively involved in the problem, sev-eral noted astronomers from other fields weighed in with comments on the possibleorigin of the DIBs (Merrill 1934). Henry Norris Russell, Pol Swings, MeghnadSaha, and Otto Struve, all of them among the fathers of modern astrophysics,commented that the DIBs were likely to have molecular origins. Later, GerhardHerzberg also became interested in the problem and spent significant effort in at-tempting to identify possible molecular carriers, though nothing definitive cameof this.

In the interval between the 1930s and the 1970s, however, most astronomersinterested in the DIB problem favored a non-molecular origin. Instead of gas-phasemolecules in the interstellar medium (ISM), dust grains with impurity centers cameinto vogue. The movement away from molecular explanations was based primarilyon the good correlation of DIB strength with interstellar dust extinction, andon perceived improbability of forming and maintaining a significant populationof molecules in the diffuse ISM where the DIBs reside. Friedman et al. (2010)summarized the literature on DIB/dust extinction.

An important point to note is that none of the correlations with dust parame-ters were perfect, implying there is no simple relation between dust and DIBs.

3 Molecular origins of the DIBs

One of Herzberg’s associates, Alexander E. Douglas, published a very significantpaper suggesting that small carbon chain molecules could be the carriers, an ideasupported by other authors soon after (Danks & Lambert 1976; and Smith et al.1977).

The impetus for the return to the original idea of molecular origins for theDIBs came from several directions: (1) by the 1960s, astronomers had discoveredion-molecule chemistry (long known by then to atmospheric chemists), whose highreaction rates could explain how significant molecular abundances could persist inthe interstellar gas; (2) also starting in the 1960s, radio astronomers were begin-ning to detect rotational emission transitions from interstellar molecules in darkinterstellar clouds, lending credence to the general idea that molecules could formand survive in space; (3) in the 1980s strong evidence was found for very largeinterstellar molecules, specifically the polycyclic aromatic hydrocarbons (PAHs);(4) at about the same time, accurate linear detectors such as CCDs revolution-ized observations of the DIBs, allowing for the measurement of very precise bandprofiles and strengths, far surpassing the older photographic spectra and settling

T.P. Snow and J.D. Destree: DIB History and the UV 343

a number of issues such as the presence or absence of weak emission wings; and(5) last, but by no means least, in the 1980s scientists from other fields such aschemistry began to take notice of the DIB problem and to lend their expertise toits solution. Chemists are now among the leading users of telescopes for spectro-scopic observations of the DIBs, and astronomers have begun to collaborate withchemists in laboratory chemistry experiments on candidate DIB carriers (see thereview by Bierbaum et al. in this volume).

In short, starting in the 1980s, the DIB problem has achieved a level of celebritythat it lacked for the first 60 years following Heger’s discovery, and has risen to alevel of awareness that, we can hope, will lead to a timely resolution.

4 Properties of the DIBs

The list below mentions several properties of the DIBs that any successful assign-ment must explain.

Limited wavelength region: Most are in the red range; almost none arein the blue.

Generally weak strengths: DIBs have very small depths, there are nonebelow 20 percent down from the continuum. Most are less than 5 percentbelow continuum and less than 30 mA in equivalent width.

Specific widths: The broad DIBs are up to 20 A wide, while the narrowones are less than 1 or 2 A wide.

Cloud type dependent: Slightly different sets of DIBs are seen in differentenvironments; they are not constant everywhere. One factor seems to becloud density, from diffuse to translucent.

Lack of saturation: All DIBs, whether strong or weak, grow linearly withincreasing extinction.

Not circumstellar: Several studies of various types of stars with circum-stellar material do not show DIBs, despite many searches.

Well-correlated with extinction: All correlations with extinction aregood (meaning correlation coefficients are generally above 0.7), but they arefar from perfect. This includes UV extinction.

Well-correlated with atomic hydrogen: DIBs are significantly bettercorrelated with atomic hydrogen than with molecular hydrogen.

Poor correlations between DIBs: Different authors have defined differ-ent sets or “families” of the DIBs that seem to be well correlated. Only onepair of DIBs comes close to being perfectly correlated.

Invariant central wavelengths: DIB central wavelengths show little tono shift between sightlines, except for Doppler shifts with cloud motion.


Distinct profiles: Many DIBs are asymmetric, though some of the broadones are symmetric. The profiles are invariant in most lines of sight.

Intrinsic structure: Some narrow DIBs show structure possibly consistentwith rotational bands. This is not caused by Doppler shifts, as they are thesame everywhere. Broad DIBs generally do not show structure.

Rarely found in emission: DIBs in emission have only been observed inthe Red Rectangle and toward one other star.

Lack of emission wings: No DIB has shown emission wings in any envi-ronment.

Lack of polarization: Several DIBs have been studied in searches forpolarization, with no positive results.

Reasonable elemental composition: Expectations about ISM gas com-position should be consistent with the cosmic abundances of common ele-ments.

Lack of UV DIBs: Recent searches have shown that any UV DIBs areweak and do not resemble strong DIBs such as λ4428.

5 PAHs as DIB carriers

Neutral PAHs tend to have their spectral lines and bands in the UV, which isone reason for thinking ionized PAHs are reasonable candidates for the DIBs car-riers (Crawford et al. 1985; Leger & d’Hendecourt 1985; and van der Zwet &Allamandola 1985). The neutral/ion ratio depends in the size of the PAH – thelarger the neutral, the greater the tendency to be singly ionized and, thus, to havetheir spectra in the visible range (Leger & d’Hendecourt 1985; Bakes & Tielens1994; Salama et al. 1996; Le Page et al. 2001, 2003).

It could be that ionized PAHs are responsible for most DIBs because of theirbetter correlation with H I than with H2 (Fig. 1). In this scenario, the C2 DIBs(A set of several DIBs that correlate well with N(C2)/E(B − V ), Thorburn et al.2003) could be located deeper in clouds, where H2 dominates.

The differing physical conditions from cloud to cloud could explain the so-called“families” of DIBs. For example, despite several searches (summarized in Lunaet al. 2008), no DIBs have been found in circumstellar environments, where thephysical conditions are different from interstellar clouds.

Since PAHs are composed of common elements, the DIB composition require-ments are easily met.

The other criteria listed above apply to almost any hypothesis on DIBs origin,except those involving solid-state transitions in dust grains. Note that moleculesadhering to dust grains are still reasonably possible DIB carriers.


Fig. 1. The physical regimes of the interstellar medium. The ionization levels are shown

for hydrogen and carbon, as those species dominate, from diffuse to dense interstellar

clouds. The horizontal lines show which types of clouds host the DIBs. The shaded area

shows the region now available with the COS.

6 UV DIBs

Previous studies have failed to find convincing evidence of DIBs below 3200 A(Snow et al. 1977; Seab & Snow 1985). These searches were confined to stars nofainter than about fourth magnitude (Copernicus) to seventh magnitude(International Ultraviolet Explorer). The COS can observe reddened stars to tenthmagnitude with EB−V values of up to 1.5 magnitudes. Hence the COS has a muchbetter chance of finding UV DIBs that any previous instruments.

Details on the design and performance of the COS can be found in Greenet al. (2010, in prep.) and Osterman et al. (2010, in prep.). Figure 2 shows theextinction limits of the COS and several other UV instruments with comparablespectral resolutions. Depending on cloud type as shown by RV (which denotes thecurvature of the extinction curve, in which small values mean high UV extinctionand high values mean low UV extinction), we see that the COS is far more sensitivethan other instruments in searches for UV DIBs.

We have published one paper on UV DIBs using archival data, finding a fewfeatures (Destree & Snow 2009). These unidentified features are all weak andnarrow, nothing comparable to the 4428 A or 5780 A features. But we note thatbroad DIBs were not included in the search, because the HST spectrograph we used


Fig. 2. A comparison of several ultraviolet spectrographs, showing the flux limits (below)

and the visual extinction by magnitudes (on the vertical axis). The numbers at the top

indicate the spectral resolutions of the various instruments. With low UV extinction

(high RV values), the COS can observe dense clouds; with high extinction, the COS is

more limited, but several magnitudes more sensitive than any previous UV spectrograph.

(the STIS) is an echelle design with many spectral orders covering less than 10 Aeach. It is difficult to join orders and match them seamlessly enough to be sureabout any broad features. Also, distinguishing between DIBs and instrumentalblemishes presents a significant challenge given the current data archive.

The COS has similar problems, though not as severe. In the far UV, meaningapproximately 1100 A to 1700 A, there are two wide spectral bands whose centralwavelengths are adjustable. Wide DIBs should be seen, if present. In the near UV(∼ 1900 A to 3200 A), the spectrograph is not as efficient and does not lend itselfto a search for UV DIBs.

In a preliminary search of COS data we found a few unidentified features,which are apparently weak, but we have more work to do before we can makeany definitive conclusions. Unfortunately, we have no observation of unreddenedstars of the spectral types to use for comparison, a standard technique for findingvisible-wavelength DIBs. Comparison stars or stellar models may be necessary ifwe want to find very weak or broad UV DIBs.

7 Final comments

All the evidence we have so far fits PAHs as the DIB carriers. But PAHs are by nomeans the only molecular family that might be consistent with the observations.


It is not necessary to assume that only one family is responsible for the DIBs; itcould be a mixture of several different species. We should be careful to have openminds.

This research was supported by NASA grant NNX08AC14G.

References

Bakes, E.L.O., & Tielens, A.G.G.M., 1994, ApJ, 427, 822

Crawford, M.K., Tielens, A.G.G.M., & Allamandola, L.J., 1985, ApJ, 293, L45

Danks, A.J., & Lambert, D.L., 1976, MNRAS, 174, 57, L25

Douglas, A.E., 1977, Nature, 269, 130

Destree, J.D., & Snow, T.P., 2009, ApJ, 697, 684 (erratum: ApJ 702, 823)

Friedman et al., 2010, ApJ, submitted

Green, J., et al., 2010, ApJ, in preparation

Heger M.L., 1922, Lick Observatory Bull., 10, 337, 146

Herbig, G.H., 1995, ARA&A, 33, 19

Herzberg, G., 1955, Mem. R. Soc. Liege, Serie, 15, 326

Krelowski J., 2002, Adv. Space Res., 30, 1395




Luna, R., Cox, N.L.J., Satorre, M.A., et al., 2008, A&A, 480, 133

Merrill P.W., 1934, PASP, 46, 206

Osterman, S., et al., 2010, ApJ, in preparation

Russell, H.N., 1935, MNRAS, 95, 610

Saha, M.N., 1937, Nature, 139, 840


Sarre, P.J., 2006, J. Molec. Spectr., 238, 1

Sarre, P.J., Miles, J.R., & Scarrott, S.M., 1995, Science, 269, 674

Seab, C.G., & Snow, T.P., Jr., 1985, ApJ, 295, 485

Smith, W.H., Snow, T.P., & York, D.G., 1977, ApJ, 218

Snow, T.P., Jr., York, D.G., & Resnick, M., 1977, PASP, 89, 758

Snow, T.P., 2002, Spectrochimica Acta Part A, 57, 615

Swings, P., 1937, MNRAS, 97, 212

Thorburn, J.A., et al., 2003, ApJ, 584, 339

van der Zwet, G.P., & Allamandola, L.J., 1985, A&A, 146, 76


THE PAH-DIB HYPOTHESIS

N.L.J. Cox1

Abstract. In this contribution I review the PAH-DIB hypothesis.Firstly, I list several properties of PAHs and their behaviour and ex-pected signatures in space. Next, I give an overview of the currentobservational insights on the DIB carriers. I conclude with a brief de-scription of recent results on the search for the DIB carriers and theprospects we have for identification.

1 Introduction: The diffuse interstellar band problem

The discovery of stationary absorption lines toward spectroscopic binariesprompted the discovery of the interstellar medium (Hartmann 1904). In the 1930san increasing number of the narrow absorption bands were identified with e.g. ion-ized titanium, neutral potassium, CH and CH+. However, a handful of broaderbands could not be identified. Merrill et al. (1938) presented the first detailedstudy of several of these so-called “diffuse” interstellar bands (DIBs). They foundthese absorption features to correlate roughly with the amount of interstellar dustinferred from the apparent reddening of stars. Also, they found a good correlationwith atomic gas phase species. Based on their observations they concluded thatindeed these DIBs could not be due to atoms but their carriers would likely beeither related to grains or to larger gas-phase molecules (see also Snow & Destree,elsewhere in this volume). Despite many efforts over the last decades, to this datethe carriers of the DIBs remain an illusive component of the diffuse interstellarmedium (ISM).

2 The PAH-DIB hypothesis

In the May 1985 issue of A&A both Van der Zwet et al. and Leger et al. stipulatedexplictly the so-called PAH-DIB hypothesis. Shortly afterwards, June 1st 1985,

1 Instituut van Sterrenkunde, K.U. Leuven, Celestijnenlaan 200 D, Leuven, Belgium;e-mail: [email protected]



Crawford et al. published a paper in similar vain in ApJ. The basic idea of thePAH-DIB hypothesis proposed by these authors states that:polycyclic aromatic hydrocarbons (PAHs) are predicted to be the car-riers of (some of) the diffuse interstellar bands1.

Essentially this was a direct follow-up on the pioneering work by Allamandolaet al. (1985) and Puget et al. (1985) in which it was shown that PAHs emit in theunidentified infrared (UIR) bands due to excitation upon absorption of optical-UV photons. In this contribution on the connection between PAHs and the DIBcarriers I will also consider related large molecules such as fullerenes, nanotubesand carbon rings. Next, I will highlight the different arguments for this connection.

3 Properties of PAHs and their behaviour in space

The arguments favouring PAHs as DIB carriers follow directly from the propertiesof PAHs (many of which were already listed in the three seminal papers mentionedabove):

• PAHs (or more generically aromatics) are very abundant in space as evi-denced from the ubiquitous presence in space of the UIR emission bands.

• PAHs, in particular if compact, can be very stable against photo- and thermo-dissociation. Therefore, a significant number density could survive in theharsh ISM.

• PAHs and their ions (cations or anions) have transitions in the near-UV, vis-ible to near-infrared. Wavelengths of the strongest bands depend strongly onthe size, the symmetry (shape) and ionization state of the considered PAH.Small to medium sized ionized PAHs have spectra dominated by one strongtransition. This can be a broad & strong UV/visible band or a narrow &weak visible/near-IR band, depending whether the ionized PAH is compactor not. This can explain the distribution of DIBs between broad/strong andweak/narrow bands.

• Laboratory spectroscopy of PAHs in Ne/Ar matrix isolation shows indeedstrong bands in the DIB range. However, inaccuracies in position and bandbroadening due to the matrix environment make it impossible to assign spe-cific bands in observations; Theoretical calculations support this conclusionbut also lack sufficient accuracy. Unfortunately, laboratory gas-phase spec-tra of PAHs in interstellar conditions are still very difficult to obtain (Pinoet al., elsewhere in this volume).

• Possibly many stable PAH configurations could be formed, and thus give riseto many absorption lines.

1Historical note: Interestingly, Platt (1956) and Donn (1968) already proposed “PAHs” toexplain the reddening curve. Several others proposed polyatomic molecules ”carbon chains/rings”as DIB carriers; e.g.: Merrill (1938), Danks & Lambert (1976), Smith et al. (1977), Douglas(1977). In addition, much was already known about UV/visible spectra and molecular propertiesof PAHs (e.g. Clar 1964; Birks 1970).

N.L.J. Cox: The PAH-DIB Hypothesis 351

• Diffuse cloud models show that PAH ionization and hydrogenation levels aresensitive to the local physical conditions, dust extinction properties, and theeffective radiation field (Bierbaum et al. and Montillaud et al., elsewhere inthis volume). A significant fraction of PAHs will be singly positively ionized(typical ionization potentials of PAHs are between 5 to 8 eV). This providesa natural explanation for the lack of correlation between DIBs.

• PAHs and/or related carbonaceous material can also account for the 2175 AUV bump (Mulas et al., elsewhere in this volume).

From the inferred presence of PAHs in space (through the UIR emission bands)and the particular chemical and physical properties of PAHs listed above one couldalready predict that these molecules should also give rise to electronic absorptionlines in the UV/visible/near-IR range.

4 Observational insights on the DIB carriers: Are they PAHs?

There is growing observational support to assign the DIB carriers to PAHs (and/orother closely related molecules). This section present the insights on the DIBs froman observational point of view. The reader is also referred to the excellent reviewsby Herbig (1995) and Sarre (2006) for additional details and references.

• DIBs show a large variety of profiles. Widths range from 0.5 to 30 A (or2–80 cm−1). These widths exclude atomic or diatomic carriers and suggestpolyatomics. Several narrow DIBs have a pronounced substructure. BroadDIBs are not known to show substructure. The 6613 A DIB substructure(peak separation) changes consistently with temperature. DIB linewidth andline shape can be explained by rovibronic contours and line broadening dueto internal conversion for medium sized molecules (i.e. PAHs containing20–30 carbon atoms or the C60 fullerene).

• Searches for polarisation of the DIB profiles have set stringent upper limitsto the polarization efficiency of the carriers. This is indicative of a carrierthat is not attached or part of (i.e. an impurity) the dust grains causing thecontinuum polarization.

• The signature of these bands can be observed to stunning similarity in galac-tic and extragalactic regions. DIBs have been observed to reside in mostLocal Group galaxies (e.g. LMC, SMC, M 31, M 33), as well as in more dis-tant spiral galaxies and starburst galaxies and even beyond toward DampedLyman-Alpha systems (e.g. Cox & Cordiner 2008). DIB carriers are there-fore ubiquitous (just as the UIR emission bands) throughout the Universe(in space and time). Formation routes and conditions are clearly universal.

• Most DIBs show only a marginal correlation with each other. One of the fewexceptions seems to be the 6196 A and 6613 A DIB pair. This implies thatthere is a unique carrier for each diffuse band (i.e. 1 DIB for 1 PAH; wherethe latter can be a specific – ionization/hydrogenation – state of a particularPAH structure).


• DIB strengths are only loosely correlated with the total amount of dust andgas. There appears to be one exception to the rule. The 8620.4 A “Gaia”DIB shows a strong correlation with reddening, E(B − V ), although someintrinsic scatter is still observed. Possibly its carrier is more closely related todust grains than others. DIB strengths have been compared to many otherenvironmental parameters that give information on the interstellar clouds.Positive correlations are observed for most cases although this may oftenbe simply the result of averaging several individual clouds along distantsightlines. Large variations in DIB strength are observed between singleclouds that contain similar amounts of gas and dust.

• Band strength ratios are affected by the effective UV field. In particular the5780/5797 ratio can be used as a tracer of the effective interstellar radiationfield (ISRF) strength.

• DIB families (i.e. bands that behave in a similar sense) reflect the interstel-lar conditions which determine locally the relative importance of neutrals,cations and anions.

• Until this date the DIBs are only observed to reside in the diffuse-to-translucent ISM (probed by sightlines toward stars). No DIBs have beenconclusively found to be present in other astrophysical environments. Forexample, evolved stars are known factories of dust and complex moleculesand could therefore contribute to the production of PAHs. However, recentstudies show that DIB carriers are very weak, if not absent, in circumstellarenvironments of post-AGB stars - thus the carriers do not form or conditionsto excite these transitions are not met.

5 Search for DIB carriers

The success of laboratory efforts to measure spectra of molecules in the gas-phasehas driven the search for carbon chains in the ISM. Maier et al. (2001, 2002)obtained laboratory spectra of C3 and C5 but only detected the former in a fewtranslucent clouds. In addition, the diacetylene cation (HC4H+) is proposed tomatch with the 5069 A DIB (Krelowski et al. 2010). There have been no otherpositive results yet for other neutral, cationic and anionic linear carbon-chainradicals measured in gas-phase (Tulej et al. 1998; Motylewski et al. 2000). In-creasingly sensitive astronomical spectra could well reveal these intrinsically weakbands in the future. Further progress in laboratory experiments have yieldedthe first gas-phase spectra of small PAHs. Iglesias-Groth et al. (2008,2010) found(new) absorption features, toward an interstellar region with anomalous microwaveemission, at wavelengths predicted for both naphthalene (C10H8

+) and anthracene(C14H10

+) cations. The conclude that 0.008% of the carbon budget is containedin the two molecules. Linnartz et al. (2010) studied an acetylene plasma andfound a band coinciding with the 5849 A DIB. A specific molecular assignmentof the carbon based carrier could not be made. Similar results by Reilly et al.(2007) showed that radical species produced by discharge of benzene give rise toan absorption band coincident with a strong DIB at 4760 A.

N.L.J. Cox: The PAH-DIB Hypothesis 353

One of the more tentative assignments is perhaps that of C+60 to two near-

infrared bands at 9577 and 9632 A. Early after its discovery C60 was proposedas contributing both to the UIR bands and the DIBs (e.g. Kroto et al. 1985;Leger et al. 1988). Sellgren et al. (elsewhere in this volume) report detection andassignment of C60 in the IR. Optical laboratory spectra of C+/−

60 in the gas-phaseare needed to detect this molecule in the visual range.

However, it is important to note here that despite identifying several newinterstellar features none of the over 200 diffuse bands observed in the visiblespectrum has been identified yet.

6 Prospects

• The PAH-DIB proposal provides a plausible working hypothesis to approachthe issue of identifying the DIB carrier(s).

• Alternative candidate carriers can not be dismissed, although the carriersare very likely large carbonaceous gas phase molecules that are stable, UVresistant, but sensitive to the local cloud conditions, in particular the ISRF.

• Spectroscopic signatures in both the UV and the near-IR are predicted formany (neutral/cation) PAHs. Surveys with COS/HST (see Snow & Destree,elsewhere in this volume) and X-Shooter/VLT are underway. The presenceof electronic absorption bands in the near-IR would be a direct test of thepresence of large ionized molecules in space. At present only two interstellarabsorption band detections, at 1.18 and 1.32 μm (Joblin et al. 1990), havebeen put forward. The lack of near-IR/UV signatures of PAHs could putsevere constraints on the properties of PAHs in the diffuse/dense ISM.

• It is crucial to explore environments other than diffuse/translucent clouds.In particular, circumstellar envelopes of evolved stars (“PAH” factories) areinteresting objects.

• Identification of specific PAHs in far-IR could be directly verified in the UVto near-IR (or vice versa) as discussed in Joblin et al. (elsewhere in thisvolume).

• The conditions for efficient formation of PAHs are not yet understood. Ei-ther this could occur via evolved stars or supernovae that produce the dustcontent of the Universe or, alternatively, these are produced in-situ in theISM. The formation process could either happen bottom up starting withsmall building blocks (C, CH, C2H2, etc.) or alternatively, top down, asfragments from large carbonaceous structures (very small grains).

• Weeding out impossible isomers is another key element to the PAH-DIBhypothesis. Another important point in this respect is the PAH size distri-bution. Is it dominated by small or larger PAHs, by peri- or catacondensedPAHs?


References

Allamandola, L.J., Tielens, G.G.M., & Barker, J.R., 1985, ApJS, 71, 733

Birks, J.B., 1970, Photophysics of Aromatic Molecules (John Wiley and Sons, Ltd.,London)

Clar, E., 1964, Polycyclic Hydrocarbons (Academic Press, London)

Cox, N.L.J., & Cordiner, M.A., 2008, IAU Symp., 251, 237

Crawford, M.K., Tielens, A.G.G.M., & Allamandola, L.J., 1985, ApJ, 293, 45

Danks, A.C., & Lambert, D.L., 1976, MNRAS, 174, 571

Donn, B., 1968, ApJ, 152, 129

Douglas, A.E., 1977, Nature, 269, 130

Hartman, J., 1904, ApJ, 19, 268

Herbig, G.H., 1995, Ann. Rev. Astron. Astrophys., 33, 19

Iglesias-Groth, S., Manchado, A., Garcıa-Hernandez, D.A., Gonzalez Hernandez, J.I., &Lambert, D.L., 2008, ApJ, 685, 55

Iglesias-Groth, S., Manchado, A., Rebolo, R., et al., 2010 [arXiv 1005.4388]

Joblin, C., D’Hendecourt, L., Leger, A., & Maillard, J.P., 1990, Nature, 346, 729

Kre�lowski, J., Beletsky, Y., Galazutdinov, G.A., et al., 2010, ApJ, 714, 64

Kroto, H.W., Heath, J.R., Obrien, S.C., Curl, R.F., & Smalley, R.E., 1985, Nature, 318,162

Leger, A., & D’Hendecourt, L., 1985, A&A, 146, 81

Leger, A., D’Hendecourt, L., Verstraete, L., & Schmidt, W., 1988, A&A, 203, 145

Linnartz, H., Wehres, N., van Winckel, H., et al., 2010, A&A, 511, 3

Maier, J.P., Lakin, N.M., Walker, G.A.H., & Bohlender, D.A., 2001, ApJ, 553, 267

Maier, J.P., Walker, G.A.H., & Bohlender, D.A., 2002, ApJ, 566, 332

Merrill, P.W., & Wilson, O.C., 1938, ApJ, 87, 9

Motylewski, T., Linnartz, H., Vaizert, O., et al., 2000, ApJ, 531, 312

Platt, J.R., 1956, ApJ, 123, 486

Puget, J.L., Leger, A., & Boulanger, F., 1985, A&A, 142, 19

Reilly, N.,, Kokkin, D.L., Joester, J.A., et al., 2007, Molec. Space Lab., 85

Sarre, P.J., 2006, J. Molec. Spectrosc., 238, 1

Smith, W.H., Snow, Jr., T.P., & York, D.G., 1977, ApJ, 218, 124

Tulej, M., Kirkwood, D.A., Pachkov, M., & Maier, J.P., 1998, ApJ, 506, 69

Van der Zwet, G.P., & Allamandola, L.J., 1985, A&A, 146, 76


ELECTRONIC SPECTROSCOPY OF PAHS

T. Pino1, Y. Carpentier1, G. Feraud1, H. Friha1, D.L. Kokkin2,T.P. Troy2, N. Chalyavi2, Ph. Brechignac1 and T.W. Schmidt2

Abstract. Polycyclic aromatic hydrocarbons are a class of moleculesof broad interest that has long been explored by various spectroscopictechniques. The electronic spectroscopy of these species is of particu-lar interest since it provides a framework for the understanding of theelectronic structure of large polyatomic molecules. Such studies alsoallow the systematic investigation of electronic relaxation mechanismsin large molecules. In this review, we focus on the gas-phase experi-mental work on such systems and present the latest progress. We alsounderline the challenges that remain to be tackled. A focus on the un-derstanding of the electronic relaxation pathways at work in gas-phasePAHs will also be presented, as well as their possible manifestation inspace.

1 Introduction

Polycyclic Aromatic Hydrocarbons (PAHs) are found in many diverse environ-ments on earth and in space, either from natural or industrial sources. Becauseof their implication in various physico-chemical phenomena, many studies havebeen devoted to them. This peculiar class of molecules has also received attentionfrom fundamental studies towards a better understanding of molecular structureand dynamics. A comprehensive review of their electronic structure can be foundin Clar’s series of books (Clar 1964a; Clar 1964b), where the basic PAH sciencemay also be found. A few examples illustrate the importance of PAH science andtheir profound impact on molecular science. The study of their quantum yieldof fluorescence during the sixties and early seventies (Hunt et al. 1962; Siebrand1967) led to major developments in the understanding of intramolecular dynamicsand radiationless transitions in particular (Englman & Jortner 1970). At the be-ginning of the eighties, with the advances of molecular beam techniques, the first

1 Institut des Sciences Moleculaires d’Orsay, Univ Paris Sud, CNRS, France2 School of Chemistry, The University of Sydney, NSW 2006, Australia



systematic studies of their gas-phase electronic spectra were performed. Two ma-jor breakthroughs in molecular science were thus achieved: the first high-resolutionspectrum of a large polyatomic molecule, fluorene, was measured in 1984 (Meertset al. 1984) and the first study of intramolecular vibrational energy redistribution(IVR) in such molecules was measured for anthracene using a picosecond laser(Felker & Zewail 1985; Felker & Zewail 1985b). Such studies mark the birth oftime-resolved intramolecular dynamics in large polyatomic systems.

PAHs are thought to carry at least some of the diffuse interstellar bands (DIBs)(Leger & d’Hendecourt 1985), some of the Red Rectangle emission bands (Sarre2006; Sharp et al. 2006), the aromatic infrared bands (AIBs) (Leger & Puget 1984;Allamandola et al. 1985) and may have a significant contribution to the “bump”at 217 nm (Joblin et al. 1992; Li & Draine 2001). All these spectral featuresimply electronic absorption and/or intramolecular dynamics such as electronic re-laxation and IVR. However, although PAH science may be thought as mature, theexperimental data that are of real pertinence for the identification of interstellar orcircumstellar PAHs are rather scarce. In fact, the questions posed by the existenceof these compounds in space are all real challenges for laboratory astrophysicistsand thus the “PAH hypothesis” has triggered many fundamental studies. How-ever, concerning the electronic spectroscopy of PAHs, only few cations have beenspectroscopically investigated in the gas phase (see Salama 2007 for a review), evenfewer neutral radicals and no anions. Deep-UV spectra around the bump regionat 217 nm are available only for few gas-phase neutral species (Joblin et al. 1992;Suto et al. 1992). This may seem somehow disappointing, but recent progresseson PAH science may be seen as a prelude for many advances in a relatively nearfuture.

In this review, a focus on the recent progress on the electronic spectroscopyof PAHs will be presented. This review mainly concerns works published afterthe review of Salama (Salama 2007) and is hopefully pedagogical enough to pro-vide the astrophysicist with the background to evaluate the work that has beenachieved and that remains to be done in order to make progress on testing thePAH hypothesis in the laboratory, and enable firm identification of some species.The first section will present the experimental methods that are used or developed,followed by a short tour on the main results in the second section. Conclusionsand perspectives will be thus given in a third section.

2 Experimental approaches

2.1 The constraints for laboratory astrophysics studies of the PAHs

The PAH hypothesis relies on the molecular rather than solid behavior of the car-rier of the AIBs (Puget & Leger 1989). It involves the transient heating mechanism,which may be explained within the framework of molecular physics as follows: af-ter absorption of a starlight photon through an electronic transition, electronic

T. Pino et al.: Electronic Spectroscopy of PAHs 357

relaxation takes place dominantly by nonradiative processes and subsequent IVRscrambles the resulting vibrational energy over all accessible vibrational states ofthe electronic ground state. Thus infrared fluorescence proceeds on the groundelectronic state. These AIBs imply very large molecules – up to a few hundredsof carbon atoms – completely isolated in vacuo on timescales ranging from a fewtens of femtoseconds up to a few seconds.

In order to study the astronomical PAHs in the laboratory, generally gas-phaseisolation is a strong requirement, in particular when spectroscopy or intramolec-ular processes are concerned. Thus, depending on the specific timescales of themechanism being investigated, several methods for isolation are necessary, butalso several experimental methods for interrogation of the molecules as well. Moststudies use laser technology that permits the study of ultrafast dynamics with fem-tosecond lasers or high-resolution spectroscopy with continuous lasers. Therefore,laboratory astrophysics devoted to PAHs necessitates the use of state-of-the-artmethods taken from molecular and/or cluster sciences that combine gas phasemolecules either in a molecular beam or ion trap with different laser sources anddetection methods. In addition, it is best to work on cold or even ultracold PAHsin order to avoid spectral congestion.

2.2 Gas-phase PAHs?

PAHs are not naturally available in the gas phase, except when the experimentuses a gas-phase reactor directly coupled to diagnostics. As such, they have tobe vaporised before being studied. Many approaches have needed to be developedsince a simple heated oven becomes less efficient at sample vaporization beyondthe size of coronene. The most successful method for large molecular systems islaser desorption (Haufler et al. 1991). Many different protocols may be adaptedto each case, using different wavelength, pulse duration and source geometries.

Neutral PAHs are of prime interest, but cations and radicals are also to bestudied. In this case additional ingredients have to be added to produce suchspecies. Laser ionisation (Pino et al. 1999), laser photolysis and electric discharge(Romanini et al. 1999) are favorite tools for this purpose but these additionalsteps render the experiment much more complex and only few species have beeneffectively studied.

Before spectroscopic investigation, it is better to cool the PAH. However heat-ing, desorption, ionization, and discharge are all in favor of producing hot PAHsrather than cold ones. As such, these methods are usually coupled to supersonicexpansions in order to cool down the target molecule (Haufler et al. 1991). Whenadding all these techniques, one may realize that not all has been effectively ex-perienced yet! Producing a large PAH cold in the gas phase is, in fact, the mostlimiting problem for experimental investigation, either because the PAH itself isdifficult to produce in large enough quantities or because the necessary densitiesin the gas phase for a given technique cannot be achieved. However, with the ad-vances on new production methods and with the very useful guidance of quantum


Fig. 1. A schematic view of the experimental set-up that was built at Orsay. The

combustion chamber contains the flat burner, cooled by circulating water. The flame

is horizontal and the flow (pre-mixed) is maintained at a constant pressure within the

chamber, at about a few tens of mbar. The burner is movable along the horizontal axis.

The quartz sampling cone (shown) allows species produced at a given distance from the

burner to enter the copper thermalisation chamber. This chamber is placed inside the

deposition chamber, which is maintained at a pressure of about 10−2 mbar under the

operating conditions. A window is then placed into the jet that is formed at the end of

the thermalisation chamber, where the gas passes through the nozzle. Possible extraction

through a skimmer forms a molecular beam at the entrance of a Time-Of-Flight mass

spectrometer.

chemical calculations, much progress has been recently reported (Kokkin et al.2008) which may strongly influence our knowledge of PAHs larger than coronenein the near future.

The source of PAHs itself is also of importance. Recent advances in chemicalsynthesis have led to the production of PAHs up to the C222H42 (Simpson et al.2002). In fact, as long as a target PAH is chosen, chemical synthesis is favoured.However gas-phase reactors offer the opportunity to directly produce many ofthem in the gas phase. Flame and laser pyrolysis are among the most promisingof such sources. At Orsay, we have incorporated a low pressure, flat and premixedflame that can be coupled to a time-of-flight mass spectrometer (Fig. 1). Underfavorable conditions, a large distribution of PAHs may be produced (Fig. 2). It isremarkable that this PAH distribution is very similar to that obtained in differentburning conditions (Apicella et al. 2007) and using laser pyrolysis or laser ablation(Jager et al. 2007; Jager et al. 2009).


Fig. 2. Mass spectrum measured with the experimental set-up Nanograins at Orsay. The

sampled flame was a low pressure (about 80 mbar) of premixed ethylene and oxygen with

a C/O ratio of 1.6. Laser ionization was used and the mass spectrum is dominated by

PAHs up to a few tens of carbon atoms.

2.3 The experimental techniques

2.3.1 Laser induced fluorescence

Laser Induced Fluorescence (LIF) is the first technique that made the measure-ment of electronic transitions of PAHs in the gas phase accessible. The first wasmeasured in the group of Smalley and published in 1981 (Beck et al. 1981a;Beck et al. 1981b). It concerned the naphthalene molecule in its neutral formand presented excitation and dispersed fluorescence spectra. The method is asfollows: resonant absorption of a tunable laser through an electronic transitionbetween the electronic ground state and an excited one is measured by collectingthe electronic fluorescence from the excited molecule. The fluorescence excitationspectrum is measured by scanning the laser wavelength and monitoring the totalfluorescence as a function of the wavelength. It is closely related to the absorptionspectrum, only intensity ratios may differ due to differences of the quantum yieldof fluorescence of the molecular states. Dispersed fluorescence spectra consist ofmonitoring the frequency resolved spectra of the fluorescence at a given resonantexcitation wavelength. The coupling of both techniques provides information onboth ground and excited electronic states, Franck-Condon factors, changes in ge-ometry, dynamics in the excited states and so on. Following advances in CCDtechnology led by astronomy, two-dimensional fluorescence has been developed.


This consists in measuring excitation and disperse spectra simultaneoulsy (Reillyet al. 2006). The strength is a fully allowed characterisation of both states, evenfrom a complex distribution of molecules arising from a gas-phase reactor. Time-resolved fluorescence spectra can also be measured which provides information onthe excited state dynamics, such as the contribution of both radiative and nonra-diative decay.

2.3.2 Resonant 2-colour 2-photon ionisation spectroscopy

Instead of detecting photons, ions (or electrons) can be measured in combinationwith a mass spectrometer. In this case, after excitation to a given electronicexcited state of the PAH, a second laser (or when possible the same laser) isused to photoionize the molecule, which is the basis of resonant 2-colour 2-photonionisation spectroscopy – R2C2PI (Duncan et al. 1981). The ion or electroncurrent thus traces the resonant absorption during the first tunable step. Thismethod is basically as sensitive as LIF but provides additional information whenions are detected – the mass of the species. This is of importance when dealingwith a distribution of PAHs, or unidentified species from a discharge (Reilly et al.2008), for example. If the first laser is tuned to a resonant transition, scanning thesecond provides the photoionization efficiency curve which reveals the ionizationpotential (e.g. Troy et al. 2009). When electrons are detected, the photoelectronspectrum may be used as a projection of the excited state wavefunction onto thecationic molecular states and provides information on the dynamics or the couplingbetween states in the neutral molecule (Stolow et al. 2004). These experimentsare generally performed in the time-resolved domain. A similar approach maybe performed on anions: The photodetachment rather than the photoionisationis measured, and appearance of the neutral and depletion of the anion must bemeasured if the detection does not involve the photodetached electrons (Tschurle& Boesl 2006).

2.3.3 Resonant multiphoton dissociation

In the case of the cations, the second ionization potential to access the doublycharged species is generally very high and cannot be easily reached for a R2C2PItechnique using benchtop equipment. In addition, aromatic cations are almostall non-fluorescent, i.e. the quantum yield of fluorescence is close to zero andelectronic relaxation processes are dominated by nonradiative phenomena. Re-garding such a situation, mass spectrometry appears as the privileged tool. Thesignature of the resonant absorption has to be traced through a measurable charge-to-mass ratio change, and involve a change in mass rather than a change in chargevia fragmentation of the cation. Actually PAHs are rather stable towards pho-todissociation. Thus, efficient measurement requires time for measuring slow pro-cesses and/or large pulse energies to accelerate the process. To overcome these


limitations, two techniques have been developed: either the cations fragment af-ter absorption of multiple photons (Useli-Bacchitta et al. 2010) or the departureof a loosely bound tagging atom (a rare gas atom for example) after absorptionof a single photon may be used to trace the absorption spectrum (Brechignac &Pino 1999; Pino et al. 1999). In the latter case the effect of the presence of theweakly bound rare gas atom has to be carefully investigated in order to properlyextract the spectrum of the bare cation. However it has been shown to be themost productive method (Salama 2007) up to now, but the latest developmentsseem promising to explore and provide novel results on cations (Useli-Bacchittaet al. 2010).

2.3.4 Cavity ringdown spectroscopy

Cavity ring down spectroscopy (CRDS) is the only direct absorption method thatcan be used for any species (O’keefe & Deacon 1988). However, although thesensitivity can be even better than the previously cited methods, the techniquesuffers from the absence of any two dimensional information which is needed forthe identification of the absorbing species if the spectrum is not self-sufficient (viathe complete rotational line’s assignment for example). It implies that as longas the spectrum can not be unequivocally assigned, no additional measurementsguide the attribution. But, it remains a valuable tool when the other methods areapplied, as a cross-check of the absorption spectrum. The principle is a follows: Alaser pulse is trapped in an optical cavity and the ringdown time is maximum anddefined by the reflectivity of the mirrors, when the laser wavelength is resonantwith a transition of the gaseous medium the ringdown time decreases due to cavityloss. Its evolution as a function of laser wavelength traces the absorption spectrum.

2.3.5 Other methods

Other experimental methods do exist and most may be considered as evolutionsand improvements upon those cited above. Either the detection or laser schemechanges, offering possibilities for peculiar aspects. Present efforts mainly concernion or electron imaging for dynamical purposes (Stolow & Underwood 2008), withthe development of detectors with time and spatial resolution, as well as opticalcavities used under various configurations (Bernhardt et al. 2010).

3 Recent progress

3.1 Toward Larger PAHs in the gas phase

Until very recently, the largest molecule conclusively identified in an extraterres-trial environment were the carbon chains HC11N, found in the circumstellar shell of


IRC+10216 by rotational spectroscopy.1 However, photophysical models of PAHsin interstellar environments conclude that they must contain at least 30 carbonatoms to be sustainable against the weathering photon flux. Prior to 2007, thelargest PAH studied as an isolated, free-flying molecule in the vacuum was ova-lene, C32H14 (Amirav et al. 1981). Recently the terrylene C30H16, which exhibitsa strong transition in the visible range of wavelength, was also studied by LIF inthe gas phase (Deperasinska et al. 2007). Even these giants are not unambiguouslylarge enough to satisfy the criteria set out by interstellar PAH models.

Troy’s all-benzenoid hypothesis motivated the study of those PAHs which canbe drawn as isolated benzene rings joined by “single” bonds. A R2C2PI and LIFspectroscopic study of the smallest of these, triphenylene (C18H12), was reportedby the Sydney laboratory in 2007 (Kokkin et al. 2007). By early 2008, an S1 ← S0

excitation spectrum of the hexa-peri-hexabenzocoronene (C42H18: HBC) was ob-tained (Kokkin et al. 2008). This R2C2PI spectrum was recorded following laserdesorption of a pressed pellet of HBC in a supersonic expansion of argon. Theskimmed beam was interrogated by a pair of laser beams some 30 cm downstreamin the extraction region of a TOF-MS. This difficult experiment stands as therecord for a PAH studied by gas-phase spectroscopy. Subsequent condensed phasespectra of HBC confirm that these α-bands near 425nm are extremely weak com-pared to bands lying in the near ultraviolet (Rouille et al. 2009). It remains forthese higher Sn ← S0 transitions to be recorded in truly isolated molecules. Theabsence of these signatures in the spectrum of the interstellar gas should provideinteresting constraints for interstellar PAH chemistry.

Larger PAHs than HBC have been synthesized. The pioneering organic chem-istry of Mullen has produced the monster C222H42 - a hexagonal sheet of grapheneconsisting of 37 separate benzenoid rings whose peripheral carbon are passivatedby H (Simpson et al. 2002). Before attempting to obtain the R2C2PI spectrumof such a beast, other all-benzenoid PAHs await, such as C78H26 and C96H30.

3.2 From the visible to the 217 nm bump region absorption of the PAHs

Electronic spectra of PAHs have been generally studied in the vicinity of the firstelectronic states (Beck et al. 1981a). Most of the gas-phase spectra of neutralPAHs were measured in the early 80’s by LIF. In the case of the cations, becauseof the DIBs hypothesis, the electronic absorption spectra were monitored in thevisible range (Salama 2007). Of course, matrix isolation spectroscopy providedimportant measurements, even up to the deep UV. Unfortunately, few were studiedin the gas phase around the bump region, most spectra being published in 1992using synchrotron light sources (Joblin et al. 1992; Suto et al. 1992). Since thattime, nearly no new spectra were published and moreover no cold ones. However

1Recently, Sellgren et al. (2010) reported the detection of the IR emission bands of C60 inreflection nebulae (see Sellgren et al., elsewhere in this volume) and Cami et al. (2010) reportedthe detection of C60 and C70 in a young planetary nebula.


it seems more and more clear that this component of the ISM should contributesignificantly to the bump at 217nm (Malloci et al. 2007).

Recently, using a gas-phase reactor for the production of large PAHs, newinsights on their electronic spectra have been obtained. At Orsay, a R2PI ex-periment coupled to a low-pressure flame has been set up. Electronic spectra ofthe small aromatics have been recorded from 207 nm up to 330 nm. Nearly 20spectra have been recorded, assigned to benzene and PAH derivatives with mono-and di-substitution by methyl, vinyl and ethynyl groups (Carpentier et al. 2010).Some correspond to known data from vapor phase measurements and others arenew. This study focused on species up to pyrene derivatives since the combustionconditions could not produce sufficient and stable enough amounts of larger PAHsfor electronic spectra to be recorded. In this experiment, the temperature of themeasured PAHs was about 300 K. In the near future, we plan to cool down theextracted PAHs from the flame using a supersonic expansion and extend the studyto larger sizes.

Important progress has been performed recently using PAHs produced by laserpyrolysis (Steglich et al. 2010). The visible to UV spectra of the molecules dis-persed directly on a subtrate in the form of thin film and in rare gas matrices weremeasured. A size dependence could be inferred. Even more the effect of conditionswere shown to be very important for the observed band profiles. The matrix iso-lation spectroscopy data contained considerably more substructures than the thinfilm. It is clear that gas-phase spectroscopy is a requirement in order to assignedproperly the carrier of the bump, even for such a broad and featureless absorption.

3.3 Ultra-high resolution spectroscopy of PAHs

During the last decade, recent advances have allowed ultra-high resolution spectraof some PAHs to be recorded. The developments were performed in the group ofM. Baba in Kyoto (Japan). Although the spectra were somehow all known, theunprecedented spectral resolution opened the possibility of a complete analysis,from spectral linewidths to rotational constants and detailed geometries. Azulene,naphthalene, pyrene and perylene, as well as benzene, have all been measured(Baba et al. 2009; Yoshida et al. 2009; Baba et al. 2009b; Semba et al. 2009;Kowaka et al. 2010). In addition, application of a magnetic field in the laser-molecules interaction zone was performed in order to probe the Zeeman effect.The outcome is a precise measure of the mixing between the first electronic S1

state with the triplet manifold. All results have revealed that in fact it is verysmall and that the nonradiative decay is dominated by the internal conversionand not the intersystem crossing.

3.4 The PAHs derivatives: Ions and radicals

As only closed shell PAHs may be purchased from chemical suppliers, these speciesdominate the spectroscopic PAH literature. However, there exist many PAHs


Fig. 3. The collection of aromatic and resonance-stabilized radicals for which excitation

spectra have been obtained by the Sydney group. The origin bands lie between 450 nm

and 600 nm. While none exhibit a perfect match to any diffuse interstellar band, the

naphthylmethyl radicals exhibit spectra in the Red Rectangle emission region (see text).

which are either open-shell or cationic, or both. All of these species have electronicspectra in the visible region, making them especially interesting as DIB candidates.Indeed, many photophysical models predict that a large proportion of interstellarPAHs should be ionized.

3.4.1 Open-shell cations

Most gas-phase spectra of cationic PAHs have been obtained about ten years ago,and have been reviewed in Salama 2007. At Orsay, a new approach is under devel-opment with the goal to simplify the photoionisation scheme for the production ofthe cations. Instead of a R2C2PI process, photoionisation is performed within thesupersonic expansion in order to produce cold cationic clusters with a single laser.Test experiments are under progress on naphthalene and its methyl derivatives.In the IRAP (ex-CESR) at Toulouse, thanks to the use of a cold ion trap, in-vestigations of the electronic spectroscopy of cations and their fragments are nowpossible and should provide challenging data on such species by resonant multi-photon dissociation of the trapped ions, as it was done on the coronene cation(Useli-Bacchitta et al. 2010).

3.4.2 PAH radicals

In the past years there have been many new spectra of polycyclic radicals reported,but almost none of radical derivatives. The significance of resonance stability in


aromatic radicals was realized with the observation of the phenylpropargyl radicalas the brightest feature, after C2, C3 and CH, in the laser-induced fluorescencespectroscopy of an electrical discharge containing benzene (Reilly et al. 2008).Phenylpropargyl is not a true PAH, but could be a precursor to PAH formation inflames and other environments (see Cherchneff 2011 in this volume), given the pu-tative importance of propargyl itself in benzene and subsequently soot formation.The radical can be considered as a benzyl and propargyl radical sharing a carbon,conferring it with a stability, as compared with methyl radical, similar to the sumof those of benzyl and propargyl. Phenylpropargyl, C9H7, which presents an ori-gin transition at 476 nm, is an isomer of the indenyl radical, never reported in thegas phase. An attempt to produce indenyl from indene yielded indanyl radical,C9H9 (Troy et al. 2009). Indene adds a hydrogen atom barrierlessly, affording theradical presenting a dominant origin at 472nm, with a true benzylic chromophore.This barrierless addition is of potential interest to interstellar chemistry, where lowreaction temperatures predominate. A benzylic chromophore is also presented by1,2,3-trihydronaphthalene, the spectrum of which was recorded by the Zwier groupat Purdue (Sebree et al. 2010). This species has its origin near 468nm, being verysimilar to indanyl radical in many respects. The Purdue group has also reportedthe spectra of 1-, and 2-hydronaphthalene, the spectra of which are dominatedby origins at 528 and 516nm respectively. Both possess extended conjugation ascompared to the benzylic chromophores. The 1-hydronaphthalene is perhaps themore interesting, it being an allylic and a benzylic radical sharing the “radical”carbon. The unligated version, Ph-CHCHCH2, which is the cinnamyl radical, wasobserved by the Sydney group very recently (Troy et al. 2010). A polycycliccinnamyl radical, was produced in Sydney by H-loss from the methyl group of2-methylindene. This radical absorbs principally at 516 nm, in the neighbourhoodof the monohydronaphthalenes.

These bicyclic radicals all possess either a benzylic (460–480nm) or a cin-namylic chromophore (510–530nm). By extending the naphthalene backbone toone extra carbon we obtain the naphthylmethyl chromophore, with absorptionsaround 580nm. The Sydney laboratory has obtained excitation and emissionspectra of both CH2-Np radicals, C11H9, and the acenaphthyl radical, a tricyclicC12H9 species. The plethora of possible radicals with this chromophore may havea relationship to the emission bands of the red rectangle nebula, which exhibitsprominent unassigned emission features in the 580nm region (Sarre 2006; Sharpet al. 2006).

The most symmetric tricyclic PAH radical is the phenalenyl radical, C13H9.Its gas-phase spectrum was recently obtained in the Sydney laboratory, it closelyresembles that obtained previously under matrix isolation. The spectrum is notorigin dominated as are those of the other aromatic radicals mentioned above.Rather, the origin is small with the spectrum being dominated by e′ vibrationmodes presumably gaining intensity through Herzberg-Teller interactions: the po-sitions are perturbed through Jahn-Teller interactions. A full analysis of thisspectrum shall be presented in the near future.


All of these open-shell PAH radicals present an opportunity to produce a dif-ferent class of PAH cation, closed-shell cations. The closed shell cations possessmuch higher HOMO-LUMO gaps than similarly sized open-shell cations. As such,their spectra in the visible region should be much sharper than the open-shellcations, and there is the possibility that they emit fluorescence. It will be possibleto measure the spectra of such cations in the near future.

3.5 Nonradiative transitions in PAHs: Present status

Nonradiative transitions are at the core of the PAH hypothesis. The main conceptsshould be recalled here. The molecular states are first described within the adi-abatic Born-Oppenheimer approximation that separates the electronic and nucleimomenta. This is done at a level beyond the harmonic approximation to accountfor Fermi resonance at low vibrational energies and wholesale intramolecular vi-brational energy redistribution higher up the well. The energy levels are obtainedand the overall mechanism may be discussed in term of time-dependent dynamicsbetween bright and dark states. It is noteworthy that such a description belongs tothe Bixon and Jortner approach (Bixon & Jortner 1968), while others prefer ma-nipulations of the eigenstates as long as it is possible, which is particularly adaptedwhen ultra-high resolution spectroscopy in the frequency resolved domain providesthe complete set of information (Quack 1990). In the case of PAHs, the dominantdescription is that from Bixon and Jortner. Electronic relaxation can then takeplace. When it occurs between two states of same multiplicity, the process re-ferred to as internal conversion (IC) and part of (or all) the electronic energy isconverted into vibrational energy. Between states of different multiplicity, inter-system crossing (ISC) occurs which relies on spin-orbit coupling. This is generallymuch slower (by about a million times) for hydrocarbons and the upper (fastest)limit should be found in the nanosecond timescale. Several configurations maybe found, depending on the shape of the potential energy surface: The change ofgeometries and the existence of surface crossings shape the radiationless processes.In the latter case, conical intersections (CI) are thought to play an important rolein intramolecular dynamics as it appears more and more than they are more sys-tematic than peculiarities (Yarkony 2001). When relaxation takes place withina particular state, only vibrational energy is concerned and the process is calledintramolecular vibrational energy redistribution (IVR). In all cases, the strengthof the coupling and the density of states (vibrational and/or electronic) controlthe nonradiative processes, there being various regimes (weak and strong couplinglimit, intermediate case and statistical limit). Very good reviews can be found inthe literature for a complete analysis of this topics (Avouris et al. 1977; Freed1978; Medvedev & Osherov 1995).

The PAH family of molecules has long served the exploration of these pro-cesses. In particular the energy gap law (EGL) was inferred from measurementof fluorescence lifetime of PAHs in solution and proved to involve ISC (Siebrand1967). The EGL states that in the statistical limit an exponential dependence of


the nonradiative rates versus the energy gap between the prepared state and theelectronic states immediately below is to be found (Englman & Jortner 1970). Asummary of nearly all available data is presented in Figure 4. Benzene derivatives,azulene derivatives, neutral and cationic PAHs are reported. Actually, althoughthe EGL was obtained from the ISC rate in neutral PAHs, it is not clear whether inIC such a dependence may be deduced here. The particular role of the CI seems tobe very important in the case of ultrafast (below 0.1 ps) electronic relaxation, andrecent calculations have shown that these intersections are systematically presentin cations (Tokmachev et al. 2010). Additionally, in benzene where a CI has longbeen known to promote IC and quench the electronic fluorescence at a certainexcess vibrational energy into the S1 state (about 4000cm−1), new insights arebecoming available and point to fast ISC as well (Parker et al. 2009). In thecase of neutral PAHs, the dominant relaxation pathway of the S1 state was longthought to be ISC, i.e. similar to the solution phase. However recent ultra-highresolution measurements of the Zeeman effect point to a weak ISC and a dominantIC (Kasahara et al. 2006; Kato et al. 2007). In fact the actual strong ground onwhich electronic relaxation of PAHs were understood is rapidly changing and weexpect a new view on this aspect in a near future.

Still, IC is clearly at work in open shell cations and this is completely consistentwith the PAH hypothesis. The existence of fast relaxation channels in highlyexcited states of neutral PAHs remains to be explored, only one study seemsto be available Suto et al. (1992). The strength of the coupling in radicals orsuperhydrogenated PAHs is still unknown.

4 Conclusions and outlook

Between our laboratories and others, there has been real progress in obtaining newspectra and sources of PAHs, their radicals and cations since 2007. In Orsay, anew experiment has been built to produce beams of large PAHs from a combustionsource, and work progresses on efficiently producing open-shell PAH cations withina molecular beam. The Sydney laboratory has obtained a collection of open-shellPAH radical spectra from electrical discharges, with the view to produce a largerdiversity of species. Looking forward, we aim to obtain close-shell PAH cationspectra in the coming years.

The astrophysical PAH hypothesis relies on the concepts of molecular non-radiative processes occuring in the gas phase. Using high-resolution spectroscopy,Baba’s group continues to uncover the links between gas phase spectroscopy andthe classical ISC studies of PAHs in solution. However, there remains a mismatchin the conditions available to laboratory spectroscopists and astronomers – that oftimescale. Unless confined to an ion-trap (where low ion numbers pose their ownproblems), it is difficult to confine a PAH to the one point in space for longer thana few microsecond – a molecule beam moves at about a millimeter per microsec-ond. However, the full mechanism underlying the PAH hypothesis takes place ontimescales up to seconds. To fully explore the dynamical behaviour of PAHs and


Fig. 4. Nonradiative electronic relaxation rate of the PAHs. Triangles are for the ISC

(intersystem crossing) measured in solution for neutral species in their hydrogenated

(full) and perdeuterated (open) form (Siebrand 1967), squares are those for IC (internal

conversion) measured neutral (azulene derivatives from Murata et al. 1972 and phenan-

threne from Ohta & Baba 1986) and cationic PAHs (for a detailed list of the references,

see Salama 2007). Circles are for the neutral fullerenes. The lines are mainly there to

guide the eyes to signify the relative efficiency of the Energy Gap Law and underline the

deuterium effect when present. In the case of the azulene derivatives, it was shown to be

very weak (Griesser & Wild 1980) unlike for the ISC in the neutral PAHs.


their cations under the full range of timescales available to astronomers, we mustperform experiments on larger scales, or slow down our molecular beams. To-wards the latter, there have been developments in Stark and Zeeman decelerationfor smaller molecules in recent years, but in favour of the former, let us go to spacewith our molecules. It would certainly be a very interesting experiment to place anartificial small comet into space and observe this with all the spectroscopic toolsavailable to the astronomical community. Such an experiment would produce awealth of data, of interest to astrochemists, astrophysicists and molecular physi-cists. Space is still the greatest laboratory of all, where so much more is possiblethan on Earth.

References


Amirav, A., Even, U., & Jortner, J., 1981, J. Phys. Chem., 85, 309

Apicella, B., Carpentieri, A., Alfe, M., et al., 2007, Proc. of the Comb. Inst., 31, 547

Avouris, P., Elbart, W.M., & ElSayed, M.A., 1977, Chem. Rev., 77, 793

Baba, M., Saitoh, M., Taguma, K., et al., 2009, JCP, 130, 134315

Baba, M., Saitoh, M., Kowaka, Y., et al., 2009, J. Chem. Phys., 131, 224318

Beck, S.M., Powers, D.E., Hopkins, J.B., & Smalley, R.E., 1981, J. Chem. Phys., 73,2019

Beck, S.M., Hopkins, J.B., Powers, D.E., & Smalley, R.E., 1981, J. Chem. Phys., 74, 43

Bernhardt, B., Ozawa, A., Jacquet, P., et al., 2010, Nat. Photon., 4, 55

Bixon, M., & Jortner, J., 1968, J. Chem. Phys., 48, 715

Brechignac, P., & Pino, T., 1999, A&A, 343, L49

Cami, J., Bernard-Salas, J., Peeters E., & Malek, S.E., 2010, Science, 329, 1180

Carpentier, Y., et al., in preparation

Cherchneff, I., present volume

Clar, E., “Polycyclic Hydrocarbons vol. I” (Academic Press and Springer Verlag 1964)

Clar, E., “Polycyclic Hydrocarbons vol. II” (Academic Press and Springer Verlag 1964)

Deperasinska, I., Zehnacker, A., Lahmani, F., Borowicz, P., & Sepio, J., 2007, J. Phys.Chem. A, 111, 4252

Duncan, M.A., Dietz, T.G., & Smalley R.E., 1981, J. Chem. Phys., 75, 2118

Englman, R., & Jortner, J., 1970, Mol. Phys., 18, 145

Felker, P.M., & Zewail, A.H., 1985, J. Chem. Phys., 82, 2975

Felker, P.M., & Zewail, A.H., 1985, J. Chem. Phys., 82, 2961

Freed, K.F., 1978, Acc. Chem. Res., 11, 74

Glinski, R.J., Michaels, P.D., Anderson, C.M., et al., 2009, Ap&SS, 323, 337

Griesser, H.J., & Wild, P.U., 1980, Chem. Phys., 52, 117

Haufler, R.E., Wang, L.S., Chibante, L.P.F., et al., 1991, Chem. Phys. Lett., 179, 449

Hunt, G.R., McCoy, E.F., & Ross, I.G., 1962, Aust. J. Chem., 15, 591

Jager, C., Huisken, F., Mutschke, H., et al., 2007, Carbon, 45, 2981

Jager, C., Huisken, F., Mutschke, H., Llamas Jansa, I., & Henning, Th., 2009, ApJ, 696,706


Joblin, Ch., Leger, A., & Martin, P., 1992, A&A, 393, L79

Kasahara, S., Baba, M., & Kato, H., 2006, Bull. Chem. Soc. Jpn., 79, 75

Kato, H., Baba, M., & Kasahara, S., 2007, Bull. Chem. Soc. Jpn., 80, 456

Kokkin, D.L., Reilly, N.J., Troy, T.P., Nauta, K., & Schmidt, T.W., 2007, JCP, 126,084304

Kokkin, D.L., Troy, T.P., Nakajima, M., et al., 2008, ApJ, 681, L49

Kowaka, Y., Suganuma, Y., Ashizawa, N., et al., 2010, J. Mol. Spec., 260, 72



Li, A., & Draine, B.T., 2001, ApJ, 554, 778

Malloci, G., Mulas, G., & Joblin C., 2007, A&A, 426, 105

Medvedev, E.S., & Osherov, V.I., 1995, “Radiationless Transitions in PolyatomicMolecules” (Springer Series in Chemical Physics), 57

Meerts, W.L., Majeski, W.A., & Van Herpen, W.M., 1984, Can. J. Phys., 62, 1293

Murata, S., Toda, T., Kokubun, H., & Iwanaga C., 1972, Ber. Bunsen-Gesellschaft Phys.Chem., 76, 1176

Ohta, N., & Baba, H., 1986, Mol. Phys., 59 921

O’keefe, A., & Deacon, D.A.G., 1988, Rev. Sci. Inst., 59, 2544

Parker, D.S.N., Minns, R.S., Penfold, T.J., Worth, G.A., & Fielding, H.H., 2009, Chem.Phys. Let., 469, 43

Pino, T., Boudin, N., & Brechignac, P., 1999, J. Chem. Phys., 111, 7337

Puget, J.L., & Leger, A., 1989, ARA&A, 27, 161

Quack, M., 1990, Ann. Rev. Phys. Chem., 41, 839

Reilly, N.J., Schmidt, T.W., & Kable, S.H., 2006, J. Phys. Chem. A, 110, 12355

Reilly, N.J., D.L., Kokkin, Nakajima, M., et al., 2008, J. Amer. Chem. Soc., 130, 3137

Romanini, D., Biennier, L., & Salama, F., 1999, Chem. Phys. Lett., 303, 165

Rouille, G., Steglich, M., Huisken, F., Henning, Th., & Mullen, K., 2009, J. Chem. Phys.,131, 204311

Salama, F., 2007, “Molecules in Space and Laboratory”, ed. J.L. Lemaire and F. Combes,Paris

Sarre, P.J., 2006, J. Mol. Spectrosc., 238, 1

Sebree, J.A., Kislov, V.V., Mebel, A.M., & Zwier, T.S., 2010, J. Phys. Chem. A, 114,6255

Sellgren, K., Werner, M.W., Ingalls, J.G., et al., 2010, ApJ, 722, L54

Semba, Y., Yoshida, K., Kasahara, S., et al., 2009, J. Chem. Phys., 131, 024303

Sharp, R.G., Reilly, N.J., Kable, S.H., & Schmidt, T.W., 2006, ApJ, 639, 194

Siebrand, W., 1967, J. Chem. Phys., 46, 440

Simpson, C.D., Diedrich Brand, J., Berresheim, A.J., et al., 2002, Chem. Eur. J., 8, 1424

Steglich, M., Jager, C., Rouille, G., et al., 2010, ApJ, 712, L16

Stolow, A., Bragg, A.E., & Neumark, D.M., 2004, Chem. Rev., 104, 1719

Stolow, A., & Underwood, J.G., 2008, Adv. Chem. Phys., 139, 497

Suto, M., Wang, X.Y., Shan, J., & Lee, L.C., 1992, J. Quant. Spec. Rad. Trans., 48, 79


Tokmachev, A.M., Boggio-Pasqua, M., Mendive-Tapia, D., Bearpark, M.J., & Robb,M.A., 2010, J. Chem. Phys., 132, 044306

Troy, T.P., O’Connor, G., Chalyavi, N., Nauta, K., & Schmidt, T.W., in preparation

Troy, T.P., Nakajima, M., Chalyavi, N., et al., 2009, J. Phys. Chem. A, 113, 10279

Tschurle, M., & Boesl, U., 2006, Int. J. Mass Spec., 249, 364

Useli-Bacchitta, F., Bonnamy, A., Mulas, G., et al., 2010, Chem. Phys., 371, 16

Yarkony, D.R., 2001, J. Phys. Chem. A, 105, 6277

Yoshida, K., Semba, Y., Kasahara, S., Yamanaka, T., & Baba, M., 2009, J. Chem. Phys.,130, 194304


SPECTROSCOPY OF PROTONATEDAND DEPROTONATED PAHS

M. Hammonds1, A. Pathak2, A. Candian1 and P.J. Sarre1

Abstract. The spectroscopic properties of protonated and deprotonatedPAHs are investigated through Density Functional Theory (DFT) cal-culations, with reference to their potential astrophysical significance.Attention is focussed on electronic and rotational spectra.

1 Introduction

Evidence for polycyclic aromatic hydrocarbons (PAHs) in the interstellar medium(ISM) is drawn almost exclusively from their infrared emission spectra – the so-called “unidentified” infrared (UIR) emission bands (see Tielens 2008, and refer-ences therein). However, it has proved to be very difficult to identify an individualPAH from these spectra as the features are indicative of molecular groups ratherthan particular molecules; observation of far-IR spectra may alleviate this prob-lem as discussed by Joblin et al. elsewhere in this volume. The spectra showsignificant variation between and within objects so if knowledge of the underlyingmix of molecules were available then profile modelling could be very productiveand act as a probe of astronomical conditions. In contrast to infrared spectra,electronic and rotational spectra are very dependent on the specific molecule withmajor differences present even between isomers. For this reason it is of interest toundertake a theoretical investigation of the electronic and rotational spectroscopicproperties of PAHs, in part as a guide for laboratory experiments. We present herepredicted electronic spectra of protonated and deprotonated PAHs, which are gen-erally closed-shell molecules and have positive and negative charge, respectively(see Fig. 1). The research seeks to contribute to the long-standing problem of thediffuse interstellar bands (DIBs) in the optical region (see Snow & Destree andCox, this volume), for which PAHs are attractive candidates. We also explore thepossibility of detection of specific PAHs through rotational spectroscopy.

1 School of Chemistry, The University of Nottingham, University Park, Nottingham, NG7 2RD,UK2 Indian Institute of Astrophysics, Koramangala, Bangalore 560034, India;Present address: Physics Department, Tezpur University, Tezpur - 784 028, India



HH

Protonated Neutral Deprotonated

+H+

-H+

Fig. 1. Illustration of the relation between neutral coronene (C24H12) and its protonated

and deprotonated forms. All three molecules have closed-shell electronic structures. In

each case the full aromaticity of the structure is maintained

2 Theoretical methods

We employ Density Functional Theory (DFT) to calculate molecular structuresand energetic properties, and its time-dependent form (TD-DFT) for predictionsof approximate electronic absorption spectra. Geometry optimisations and TD-DFT calculations were conducted using the B3LYP functional with a 6-31G* basisset for protonated PAHs and a 6-31+G** basis (Hehre et al. 1972; Hariharan et al.1973) for deprotonated (anionic) PAHs. The latter basis set is not a standard partof the Q-Chem package used in this work (Shao et al. 2006).

3 Protonated PAHs (HPAH+)

3.1 Electronic spectroscopy of protonated PAHs

Alongside research on PAHs in connection with the UIR bands, there is long-standing interest in their possible role in relation to the DIB band problem (seeHerbig 1995 and Sarre 2006, and references therein). The potential significanceof protonated PAHs, denoted here HPAH+, in connection with this problem wasfirst highlighted by Le Page et al. (1997) and both this aspect and the chemistryof HPAH+ molecules have been discussed in subsequent papers (Snow et al. 1998;Le Page et al. 2003). Although small-to-medium sized neutral PAHs generallyabsorb in the UV or near-UV part of the spectrum, and in principle would be de-tectable in absorption in the spectra of background stars, no such evidence has yetbeen found. Larger neutral PAHs such as hexa-peri-hexabenzocoronene (C42H18)with absorption bands in the 4000–4500 A range, would appear to be promisingcandidates, but to date no correspondence with DIBs has been found (Kokkinet al. 2008). Consequently attention has also been paid to ionic forms of PAHsand, partly because of ease of production in the laboratory, to radical cationswhich generally absorb in the visible spectral region. However, being open-shellmolecules, they are expected to be reactive and so alternative ionic forms including

M. Hammonds et al.: Protonated and Deprotonated PAHs 375

Wavelength (nm)

Osc

illat

or S

tren

gth

HH(a)

(b)

Fig. 2. Calculated electronic transitions of (a) protonated coronene and (b) coronene.

protonated PAHs are worthy of consideration. At present laboratory recording ofelectronic spectra of HPAH+ molecules at high-resolution is in its infancy butdeveloping, as outlined in this Symposium. For large systems the only currenttheoretical method that lends itself to predicting electronic spectra is TD-DFTand the precision is low, being in the region of ± 0.3 eV. However, the results canstill be very useful in identifying potential candidates. The calculated electronicabsorption spectrum of e.g. protonated coronene differs markedly from its corre-sponding neutral parent molecule as seen in Figure 2. Although the protonatedmolecule has transitions in the near-UV, as is the case for the neutral molecule, italso has strong transitions in the visible part of the spectrum. Although the effectof protonation on the absorption spectrum of benzene has been known for manyyears, extension to polycyclic aromatic hydrocarbons is new. The precision ofTD-DFT calculations is far too low to discuss any meaningful attribution of DIBsto specific molecules, but the semi-quantitative result is significant. In particularit may assist in planning laboratory experiments in the condensed (inert gas ma-trix) or, preferably, gas phase. The dissociation energies of protonated PAHs arelower than their neutral counterparts which does raise some question as to theirphotostability, although for large protonated PAHs fragmentation is expected tobe reduced due to the very high density of states.


Protonated CoroneneB3LYP/6-31G*

C24H13+

CoroneneB3LYP/6-31G*

C24H12

Coronenyl RadicalB3LYP/6-31G*

C24H12•

Coronenyl AnionB3LYP/6-31+G**

C24H11–

HH

Wavelength (nm)

Osc

illat

or S

tren

gth

Fig. 3. Calculated transitions of protonated, neutral, radical and deprotonated coronene.

4 Deprotonated PAHs (PAH−n−1)

As far as we are aware, deprotonated PAHs have not been considered explicitlyin an astrophysical context. The analogous deprotonated polyacetylenes such asC6H− are the subject of intense current activity as recorded through their labora-tory rotational spectra and corresponding studies by radioastronomy (McCarthyet al. 2006). PAH anions have been discussed in connection with interstellar chem-istry and as possible contributors to the UIR bands, but generally in terms of theradical anion form (Bakes & Tielens 1998). The electron affinities of closed-shellneutral PAHs (Malloci et al. 2005) are a factor of 2–3 lower than those of theneutral radicals of interest here. We have computed using DFT the electron affini-ties of a number of radicals (values in eV, for the most stable isomer) as follows:pyrenyl (1.57), coronenyl (1.68) and ovalenyl (2.00). While these electron affinitiesare lower than for the polyacetylenic radicals they are still reasonably high anddeprotonated anions should form readily through radiative electron attachment tothe radical. Given the special nature of anions a different basis (6-31+G**) wasemployed. In all cases it is found that the negative charge is very localized at thesite of proton removal.


Fig. 4. Computed pure rotational spectrum of the coronenyl anion for J values up to

100 and at 30 K. The inset shows a ∼1 GHz region of very nearly equally (harmonically)

spaced lines.

4.1 Electronic spectroscopy of deprotonated PAHs

An example of a calculated electronic spectrum, for deprotonated coronene, isgiven in Figure 3 (lowest panel) where it is compared with corresponding spectraof the protonated, closed-shell neutral and neutral radical forms. These are allover-simplifications given that no vibrational band structure is included. The de-protonated coronene molecule exhibits just one significant band in the 300–750 nmregion. The astronomical importance of this spectrum is uncertain because in dif-fuse and translucent clouds it is probable that photodetachment of the electronwould destroy the anion; the calculated band at e.g. ∼420 nm corresponds toa photon energy of ∼3 eV which is well above the ionization energy of ∼1.7 eVand so autodetachment would occur on photoexcitation. However, in denser re-gions this restriction is lifted. The possibilities of detection of such species byradioastronomy is now considered.

4.2 Rotational spectroscopy of deprotonated PAHs

The wealth of information gleaned about dense clouds through radioastronomy isreliant on the detected molecules possessing permanent electric dipole moments.


PAHs are very likely present in some form but as the simplest neutral systemsare not polar no detection is possible through their rotational spectra. One mod-ification, and one subjected to a radioastronomical search, is corannulene whichhas a significant dipole moment as it is non-planar (Pilleri et al. 2009). PAHscontaining heteroatoms may also offer some opportunities. For PAH ions it islikely that protonated forms are present but these molecules are asymmetric topsand so the spectra are extremely complex with huge line-dilution. An alternativethat might allow progress is the class of deprotonated PAHs. The removal of oneperipheral hydrogen nucleus does affect slightly the centre-of-mass of an other-wise symmetrical PAH, but does not introduce asymmetry to the extent causedby off-plane hydrogens in protonated versions. Deprotonated PAHs are thereforevery nearly symmetric top molecules but, crucially, possess large electric dipolemoments which fall in the 10–14 Debye range for the coronenyl, ovalenyl, circum-pyrenyl and cirumcoronenyl anions. Figure 4 shows a spectrum computed usingthe pgopher package (Western 2009) where at 30 K there are a large number ofapparently barely resolved lines. However, the molecules are very nearly symmet-ric tops and so the main pattern is a series of equally spaced lines as the finer(K) structure is unresolved (see inset). While the capacity to calculate accuratepredicted line positions for deprotonated PAHs is certainly beyond any currenttheoretical method, identification of harmonically related frequencies has playeda significant role in identifying linear molecules in astronomical environments, in-cluding C6H−, and these could be sought in radio data.

We thank EPSRC (MH) and STFC (AC) for studentships, the Symposium organizers and theDepartment of Science and Technology, New Delhi, for financial assistance (AP), The RoyalSociety for travel support (PJS), and The University of Nottingham for support and use of theHPC facility.

References


Hariharan, P.C., & Pople, J.A., 1973, Theoret. Chim. Acta., 28, 213

Hehre, W.J., Ditchfield, R., & Pople, J.A., 1972, J. Chem. Phys., 56, 724

Herbig, G.H., 1995, ARA&A, 33, 19

Kokkin, D.L., Troy, T.P., Nakajima, M., et al., 2008, ApJ, 681, L49

Le Page, V., Keheyan, Y., Bierbaum, V.M., & Snow, T.P., 1997, J. Am. Chem. Soc.,119, 8373


Malloci, G., Mulas, G., Cappellini, G., Fiorentini, V., & Porceddu, I., 2005, A&A, 432,585

McCarthy, M.C., Gottlieb, C.A., Gupta, H., & Thaddeus, P., 2006, ApJ, 652, L141

Pilleri, et al., 2009 MNRAS, 397, 1053


Sarre, P.J., 2006, J. Mol. Spectrosc., 238, 1

Shao, Y., et al., 2006, Phys. Chem. Chem. Phys., 8, 3172

Snow, T.P., Le Page, V., Keheyan, Y., & Bierbaum, V.M., 1998, Nature, 391, 259


Western, C.M., PGOPHER, a Program for Simulating Rotational Structure (version6.0.111), University of Bristol, http://pgopher.chm.bris.ac.uk


OBSERVATIONS OF INTERSTELLAR CARBONCOMPOUNDS

E. Dartois1

Abstract. Infrared absorption and emission features observed spectro-scopically in our Galaxy allow to probe the composition of solid dustgrains, their evolution and thus follow the cycling of matter in theGalaxy. Many observables do reveal the presence of large amounts ofcarbonaceous particles in space, other than the PAH-like emission lines.The carbonaceous materials observed include amorphous carbons, di-amondoids showing in emission for a few specific sources, and the re-cently detected fullerenes. An important hydrogenated amorphous car-bon component (HAC or a-C:H), traced by the 2940 cm−1 structuredabsorption feature is observed against Galactic background sources.Since the discovery of this feature in the early eighties (Allen 1981),the observation of a-C:H has been extended to the mid-infrared byspace observatories, giving insight into additional associated features.They are also observed in external galaxies, showing the ubiquitous na-ture of these components. We will focus on astronomical observationsof organic matter other than PAHs, amorphous carbons and associatedlaboratory dust analogues relevant to astrophysical applications.

1 Introduction

The nowdays observed dust results from the superposition and mixing of manydifferents galactic environments that participate to the so-called “dust lifecycle”.Dust is produced in several source environments and, as it travels through thegalaxy, experiences various physical and chemical processes that will modify itsstructure and/or composition. Among carbonaceous contributors, additional car-bon allotropes, other than the polycyclic aromatic hydrocarbons (PAHs) describedabundantly in this volume, participate in the building of the dusty interstellarmedium. Organic matter includes various observed structural forms such as amor-phous carbons, diamond, hydrogenated amorphous carbons and other materialswith aromatic/aliphatic mixed structures.

1 Institut d’Astrophysique Spatiale, UMR-8617, Universite Paris-Sud, Batiment 121,91405 Orsay, France



2 Amorphous carbons

Dust composition in the vicinity of evolved stars flows is constrained from directobservations and modelling. When the stellar wind flowing out reaches a pressurestill sufficiently high (10−5− 10−3 Pa) and temperature sufficiently low (∼ 2500−1000 K), atoms and molecules nucleate small dust seeds that start to grow inthe expanding flow, until they reach a region where the pressure and temperatureregime does not allow to pursue an efficient grain growth (e.g. Patzer et al. 2003;Gail & Sedlmayr 1999). This dust formation window is the consequence of thecompetition between the dust formation timescale and the stellar wind ejectionspeed that freezes the reactivity at some point. Passed this point in the flow, thedust formed leaves the nucleation regime and enters the evolution one. The kind ofdust formed is controlled by chemical parameters such as the carbon/oxygen ratioin the flow. The observed circumstellar dust near evolved stars is a mix of oxygenand carbon rich dust. Of particular interest for the carbonaceous world are the so-called “carbon” stars, injecting important amounts of carbon under an amorphousform difficult to observe remotely, as no specific features are associated with them,unlike silicates. This allotrope is clearly observed around extreme carbon stars, asa broad featureless emission in equilibrium with the radiation field (See Fig. 1).

Fig. 1. Emission spectra of two carbon stars extracted from the Infrared Space Observa-

tory database (http://iso.esac.esa.int/ida/). They illustrate the almost featureless

continuum emission by highly carbonaceous grains in evolved carbonaceous stellar flows

(see also Volk et al. 2001; Hony 2002; Chen et al. 2010 and citations therein).

These stars are likely to be the progenitors of C-rich protoplanetary nebulae(PPNs) such as AFGL 2688 (classified as a Class C by Peeters et al. 2002; see alsoPeeters in this volume), and C-rich planetary nebulae (PNs) such as NGC 7027.They will be involved in the carbon dust cycle although the grains they producedare more difficult to observe in the later phases.

E. Dartois: Observations of Interstellar Carbon Compounds 383

3 (Nano-)diamonds

The presence of high amounts of (nano-)diamonds, up to several thousands ppm(e.g. Ott 1993; Anders & Zinner 1993), in the organic extracts from meteoriteshas led to the search for their plausible infrared signatures in the laboratory, toconfront their spectra to those of young stellar objects observed in early evolutionphases eventually leading to a protosolar system. They were initially proposed asan alternative explanation for some of the unidentified infrared bands (UIBs) andprobably responsible for the “21 microns” astronomical feature, following infraredspectroscopy studies of “Allende” and “Orgueil” meteorite nanodiamonds (Koikeet al. 1995; Hill et al. 1998).

The unambiguous detection of specific CH methine and methylene stretchingmodes, observed at 3.53 and 3.43 μm, associated with hydrogen atoms chemicallybounded to diamond structure was demonstrated by Guillois et al. (1999), bycomparing emission spectra of astronomical sources to laboratory spectra of nano-crystal diamond films exposed to hydrogen by Chang et al. (1995). These sourcesare different from the ones previously discussed for the 21 “microns” feature. Re-cently, from a molecular approach, Pirali et al. (2007) measured small diamondoidsin the gas phase and, reported calculations, as well as attenuated total reflectancespectra for higher diamondoids. Analogues of the measured species with the samesymmetry containing around 130 C atoms (e.g., C136H104) show an intensity ratioof the 3.53 and 3.43 μm bands close to that observed in the interstellar spectra.The inferred size is close to typical sizes of meteoritic diamonds.

Observationally, a survey of 30 Herbig Ae/Be stars, the spectral type of thepreviously observed stars where the features had been observed, was conductedby Acke & van den Ancker (2006). No new source could be added to the list ofdiamond emission features, implying that less than 4% of the Herbig targets showthe prominent emission features at 3.43 and/or 3.53 μm.

In an effort to resolve spatially the location of these features, Habart et al.(2004) marginally resolved the emission intensity profile, showing that it is lessextended than the associated 3.3 μm PAH feature. Goto et al. (2009), using anadaptive optics system, resolved spatially the PAH and diamond CH emissions.The diamond emission is concentrated, peaking at about 30 AU from the star.The PAH emission is more extended (above 100 AU). The authors speculate ondiamond formation in these circumstellar disks, with graphitic grains processedinto diamond grains under highly-energetic particle bombardment.

4 Fullerenes

Fullerenes and in particular C60 have been searched for intensively via electronicand infrared specific transitions (e.g. Foing & Ehrenfreund 1994; Fulara et al.1993; Moutou et al. 1999; Herbig 2000). C60 was recently detected by Sellgrenet al. (2009, 2010) and Cami et al. (2010). More details can be found in thecontribution by K. Sellgren et al. in this volume.


5 Hydrogenated amorphous carbons

Hydrogenated amorphous carbons, HAC or a-C:H for amorphous material madeof C and H, constitute one important component of interstellar dust. They were ob-served initially at 3.4 μm against a Galactic center source (Allen &Wickramasenghe 1981). The features contributing to this absorption band wereearly associated to sp3 CH3 and CH2 stretching modes, as stated in e.g. Duley &Williams (1983), “Interstellar amorphous carbon dust with chemisorbed CH2 andCH3 groups may be a significant component of interstellar dust in diffuse clouds”.Since then, numerous experiments/observations have been performed to constrainits origin (e.g. Jones et al. 1983; Butchart et al. 1986; Mc Fadzean et al. 1989;Ehrenfreund et al. 1991; Sandford et al. 1991; Sandford et al. 1995; Pendletonet al. 1994; Tielens et al. 1996; Geballe et al. 1998; Chiar et al. 2002; Mennellaet al. 2002; Pendleton & Allamandola 2002). A Galactic abundance for a-C:H hasbeen estimated from observed CH stretching modes. Depending on the assumedmaterials, the associated oscillator strength for CH modes and the degree of hydro-genation, the implied cosmic carbon fraction varies from 2.6% to 35% (Sandfordet al. 1991), above 2.5 to 4% based on the spectra of alkanes (Pendleton et al.1994) and up to 20–30% for laboratory analogues of a-C:H (Duley 1994; Duleyet al. 1998).

The constrain on a-C:H was initially set only by the CH stretching modes.Nevertheless, the band profiles of many laboratory analogues were already incom-patible with the astronomical absorption profile. It is incompatible with H2O-dominated photolytic or radiolysis residues obtained starting with an interstellarice mantle composition (Bernstein et al. 1995; Allamandola et al. 1988; Greenberget al. 1995; Pendleton & Allamandola 2002). Instead, laboratory analogues usingheated carbon rod (Schnaiter et al. 1998), laser desorbed carbon (Mennella et al.1999), plasma deposition (Lee & Wdowiak 1993; Furton et al. 1999) or photo-produced a-C:H at low temperature (Dartois et al. 2005) provide a much betterfit to the observed features. The Insoluble Organic Matter (IOM) extracted frommeteorites also provides a reasonable fit to the astronomical observations (e.g.Ehrenfreund et al. 1991) at 3.4 μm.

The discrimination became clearer with the advent of space telescopes, allow-ing to track the mid-infrared fingerprints of this material. The moderate to highoxygen content in many of the laboratory analogues or IOM does provide strongabsorptions in the mid-IR fingerprint region, and the carbonyl band is one of thestrongest in the 1650 to 1800 cm−1 range. Within the signal-to-noise achieved,these absorptions can at most be a minor contribution to the observed diffusemedium spectra, as shown in Figure 2. For Galactic sources there still exists con-fusion in the mid-infrared, due to foreground dense clouds ice absorptions (Fig. 2upper spectra, Dartois et al. 2004), or local circumstellar contributions decoupledfrom the diffuse medium features (Fig. 2 lower spectra, Chiar & Tielens 2001;Chiar et al. 2002). However, meaningful upper limits were already set.

In the last years, several extragalactic obscured active galactic nuclei (AGN)sources displaying a-C:H absorptions have been observed, mainly via the stretching


Fig. 2. Continuum extracted transmittance spectra along two Galactic center lines of

sight. Both observations supposedly probe the same diffuse interstellar medium absorp-

tions, but display many additional contributions arising from circumstellar and/or dense

cloud contributions. Nevertheless, they constrain the mid-infrared counterpart of the

3.4 μm absorption from a-C:H. See text for the detailed analysis of these differences.

modes (Pendleton et al. 1994; Mason et al. 2004; Dartois et al. 2004; Imanishi2006; Risaliti et al. 2006; Imanishi et al. 2008). In NGC 1068, the hydrocarbondistribution has even been imaged to better constrain its spatial distribution, andforesee its nature (Mason et al. 2006; Geballe et al. 2009). The extragalacticsources provide a different constrain to get insight into the a-C:H structure andubiquity. Contrary to Galactic embedded sources used as infrared backgroundto probe the diffuse medium, obscured AGNs probe large dust column densitiesin front of an extended infrared continuum provided by the active nucleus envi-ronement. With a parsec-scale infrared probing pencil, the risk of very local cir-cumstellar contamination is minimised . With a favorable AGN nucleus to diffusemedium alignment geometry, a line-of-sight free from significant dense cloud iceabsorptions can be observed, providing a clean mid-infrared continuum to probea-C:H fingerprints. Another advantage of the extragalactic case is the moderategalaxy red-shift, shifting the wavelength frame to more favourable atmosphericwindows for ground-based observations, thus giving a unique access to the aro-matic CH stretching mode region, devoid of strong telluric contamination, whichare especially important for shallow absorption features.

The IRAS 08572+3915 AGN is such a test case source, with a redshift ofz∼0.0583. An upper limit on the aromatic C-H stretch has been evaluated, andthe aromatic versus aliphatic C-H content of interstellar a-C:H can be better con-strained (N(CH aromatic) / N(CH sp3) ≤ 0.08; Dartois et al. 2007). Using the


N(H) column density relation derived from the observed silicate optical depth com-bined to an elemental cosmic carbon abundance, about 15% of the cosmic C mustbe involved in sp3 bonding. The methyl and methylene C-H observed ratio implyalso x(H) / x(C sp3) ∼ 2.33. Taking into account that only the infrared moder-ately active modes are observed, this sets a lower limit to the hydrogen contentof the material observed, meaning a fractional hydrogen content in the a-C:H ofx(H)≥0.2. These constraints can be brought together in a ternary phase diagramwhere the hydrogen content and the two main bonding types for carbon (sp2 andsp3, as the sp contribution is expected to be small) constitute the poles (Ferrari& Robertson 2000; Dartois et al. 2007). Overplotted are the constraints for IRAS08572+3915, as well as the position for different laboratory produced analogues,including soot, and PAHs lying on the H to sp2 border. Random Covalent Networkmodels for materials made with olefinic or aromatic C = C, using the model de-velopped by Angus & Jansen (1988) and extended to aromatic structure by Jones(1990) are displayed. The a-C:H analogues fit perfectly with these constraints aswell as the overall infrared spectrum.

In the model structure proposed by Pendleton & Allamandola (2002, Figure 17in this paper), the aromatic CH component (∼160 by number) represents morethan two times the aliphatic CH component (CH2+CH3∼70), for a carbonaceousbackbone of ∼540 C atoms. We would expect in principle, with such a structure,an aromatic CH absorption higher than observed for IRAS 08572+3915, even ifthe lower oscillator strength of the CH aromatic stretch is taken into account.

Fig. 3. Ternary phase diagram for analogues made principally from C and H. The con-

strain from observations of IRAS 08572+3915 is set (min H and max H contents). Several

domains for various synthetic materials from the laboratory are shown. Two Random

Covalent Network models (Angus & Jansen 1988; Jones 1990) are also displayed, almost

enclosing the a-C:H domain.


Based on the observations and ternary diagram, we propose to revise the basicstructural unit (BSU) issued from the Pendleton & Allamandola analysis by using amaterial with a lower aromatic CH contribution1, such as hydrogenated amorphouscarbon or hydrogen rich soot like analogues.

Interstellar a-C:H materials remain difficult to observe as compared to PAHs.Their astrophysical detection requires large column densities, as shown by thevisual to band extinction relations (AV/τ(3.4 μm) ∼ 250; e.g. Sandford et al.1995; AV/τ(6.85 μm) ∼ 640; Dartois & Munoz Caro 2007). They represent amajor ISM component, containing up to 30% of total galactic ISM carbon (typical∼ 15%). Their spectroscopic absorption feature contrast is easily hidden by PAHswhen energetic photons are present, as strongly emitting PAH regions requireonly few tenth of an AV to be able to record bright emission bands. This explainswhy their observation is hampered in e.g. the interstellar media of Starburstgalaxies. It is ubiquitous in the Galactic diffuse ISM and observed in a largenumber of external galaxies (e.g. Dartois & Munoz-Caro 2007). It contains lowamount of O heteroatoms. It is in agreement with laboratory plasma produced(microwave, laser) and/or photoproduced a-C:Hs. It may be partly responsible forISM fluorescence features (e.g. Godard & Dartois 2010 and references therein).It is a possible precursor for another form of interstellar carbon (PAHs) uponenergetic processing.

6 Mixed aliphatic/aromatic structural units

6.1 Observed classes A, B, C

The relative profile variations of the mid-IR emission among various astrophysi-cal sources has been underlined since a long time from ground-based observations(Joblin et al. 1996; Geballe et al. 1997) and a classification in various classeswas performed when space missions provided a full wavelength infrared coverage(Peeters et al. 2002; van Diedenhoven et al. 2004, see Peeters elsewhere in thisvolume). Some of the observed sources, the so-called ”class C”, display both aro-matic and aliphatic emission bands, questioning us on the possible link betweenhydrocarbons seen in absorption and emission. In few instances, the C-H duststretching mode is spatially resolved in emission by adaptive optics (Goto et al.2003; Goto et al. 2007) and relative variations from the central star suggest a pro-gressive disappearance of aliphatic character to the benefit of the aromatic featurewhen the distance from the star increases. It has been shown that laboratory ana-logues thermally annealed at different temperatures, provide a reasonable matchto this process (e.g. 2005, 2000, 1997, 1990). However one should take into ac-count the possibility to have different emission mechanisms between both phases(solid and molecular), that will depend also on dust particles size in the stellarfield. One must not forget that concomittant mechanisms are at work (chemical

1Note that the constrain is on the aromatic CH. An aromatic backbone structure can still bepresent but interconnected to other elements via e.g. methylene or carbon-carbon bonds like ina-C:H or soot.


modification, emission, stellar versus ISRF photon irradiations) that still need tobe further constrained.

A progressive evolution among sources from class A to C has been notablyunderlined by Szczerba et al. (2005). The aromatic C = C stretching mode posi-tion is shown to shift from 6.2 to 6.3 μm while the so-called “7.7” feature evolvestoward 8 μm. The change in the 6.2 μm feature position has been associated withvarious hypothesis, such as the formation of PAH metal complexes, the formationof PAH clusters, isotopic shifts, size changes or induced by heteroatoms like nitro-gen. In the last case, the band would shift with N atom substitution site withinthe aromatic structure (Hudgins et al. 2005). In the following a simple alternativescenario is proposed to this evolution, based on recent laboratory experiments.

Fig. 4. Upper panel: infrared absorption spectra of laboratory soot samples with varying

aliphatic to aromatic C-H stretch content and associated C = C mode region. Lower

panel: evolution of the C = C position in the various samples produced as a function of

the aromatic to aliphatic ratio observed in the stretching mode region. Note that the

total H/C vary among the samples. The expected astrophysical classes range of 6.2 μm

positions are shown for comparison.

6.2 A possible aromatic C = C shift induced by aliphatics

The sources with the “6.2 μm” feature lying at 6.3 μm also display additionalemission lines around 6.85 and 7.25 μm, as well as a more or less pronounced


3.4 μm emission band on the side of the aromatic feature. Sloan et al. (2007)had proposed that the observed wavelength shifts are consistent with hydrocar-bon mixtures containing both aromatic and aliphatic bonds. More recently, Ackeet al. (2010) showed the same behaviour with a large set of Spitzer spectra (seealso Acke, elsewhere in this volume). In order to understand carbon dust nucle-ation, and the observed astrophysical classes of materials, laboratory soot analoguespectra have been recorded for 50 distinct samples, using a premixed low pressureflame. Varying the flame combustion regime with respect to the stoechiometricconditions, and forcing various sooting conditions allow to produce many differentcarbonaceous soot materials (Pino et al. 2008). The resultant analogues are inves-tigated by FTIR spectroscopy. This spectral analysis, applied to the band shapeand position variations, was then used to interpret the diversity and evolution ofthe features in the astronomical spectra. Strong correlations were evidenced be-tween the spectral regions characteristic of the C = C and C-H modes in theseanalogs. These shed light on the origin of the infrared emission features. In par-ticular, the observed shift in the position of the 6.2–6.3 μm band is shown to bea key tracer of the evolution of the aliphatic to aromatic component of carbona-ceous dust. The produced materials have been characterised by other techniques(Raman spectroscopy, TEM, elemental analysis). These analyses suggest thatthere is a structural modification of the carbonaceous backbone by aliphatics. Theevolution between these extreme positions (6.2 to 6.3 μm) thus appears as tracingthe underlying carbon skeleton structure, linked to the presence of aliphatics inthe network that favours the development of an alternative basic structural unitas compared to classical aromatic compounds.

References

Acke, B., & van den Ancker, M.E., 2006, A&A, 457, 171


Allamandola, L.J., Sandford, S.A., & Valero, G.J., 1988, Icarus, 76, 225

Allen, & Wickramasenghe, 1981, Nature, 294, 239

Anders, E., & Zinner, E., 1993, Meteoritics, 28, 490

Angus, J.C., & Jansen, F., 1988, J. Vacuum Sci. Technol., 6, 1778

Bernstein, M.P., Sandford, S.A., Allamandola, L.J., Chang, S., & Scharberg, M.A., 1995,ApJ, 454, 327

Butchart, I., McFadzean, A.D., Whittet, D.C.B., Geballe, T.R., & Greenberg, J.M.,1986, A&A, 154, L5

Chang, H.-C., Lin, J.-C., Wu, J.-Y., & Chen, K.-H., 1995, J. Phys. Chem., 99, 11081

Chen, P.-S., Yang, X.-H., & Shan, H.-G., 2010, Res. Astron. Astrophy., 10, 363

Chiar, J.E., & Tielens, A.G.G.M., 2001, ApJ, 550, L207

Chiar, J.E., Adamson, A.J., Pendleton, Y.J., et al., 2002, ApJ, 570, 198

Dartois, E., Marco, O., Munoz-Caro, G.M., et al., 2004, A&A, 423, 549

Dartois, E., Munoz Caro, G.M., Deboffle, D., Montagnac, G., & D’Hendecourt, L., 2005,A&A, 432, 895


Dartois, E., et al., 2007, A&A, 463, 635

Dartois, E., & Munoz-Caro, G.M., 2007, A&A, 476, 1235

Duley, W.W., & Williams, D.A., 1983, MNRAS, 205, 67P

Duley, W.W., Scott, A.D., Seahra, S., & Dadswell, G., 1998, ApJ, 503, L183

Duley, W.W., 1994, ApJ, 430, L133

Ehrenfreund, P., Robert, F., D’Hendecourt, L., & Behar, F., 1991, A&A, 252, 712

Ferrari, A.C., & Robertson, J., 2000, Phys. Rev. B, 61, 14095

Foing, B.H., & Ehrenfreund, P., 1994, Nature, 369, 296

Fulara, J., Jakobi, M., & Maier, J.P., 1993, Chemical Physics Letters, 211, 227

Furton, D.G., Laiho, J.W., & Witt, A.N., 1999, ApJ, 526, 752

Gail, H.-P., & Sedlmayr, E., 1999, A&A, 347, 594

Geballe, T.R., 1997, From Stardust to Planetesimals, 122, 119

Geballe, T.R., Chiar, J., Pendleton, Y.J., & Tielens, A.G.G.M., 1998, Astrophys. SpaceSci., 255, 457

Geballe, T.R., Mason, R.E., Rodrıguez-Ardila, A., & Axon, D.J., 2009, ApJ, 701, 1710

Godard, M., & Dartois, E., 2010, A&A, forthcoming

Goto, M., Maihara, T., Terada, H., Kaito, C., Kimura, S., & Wada, S., 2000, Astron.Astrophys. Suppl., 141, 149

Goto, M., et al., 2003, ApJ, 589, 419

Goto, M., et al., 2007, ApJ, 662, 389

Goto, M., et al., 2009, ApJ, 693, 610

Greenberg, J.M., Li, A., Mendoza-Gomez, C.X., et al., 1995, ApJ, 455, L177

Guillois, O., Ledoux, G., & Reynaud, C., 1999, ApJ, 521, L133

Habart, E., Testi, L., Natta, A., & Carbillet, M., 2004, ApJ, 614, L129

Herbig, G.H., 2000, ApJ, 542, 334

Hill, H.G.M., Jones, A.P., & D’Hendecourt, L.B., 1998, A&A, 336, L41

Hony, S., 2002, Ph.D. Thesis

Hudgins, D.M., Bauschlicher, C.W., Jr., & Allamandola, L.J., 2005, ApJ, 632, 316

Imanishi, M., 2006, AJ, 131, 2406

Imanishi, M., Nakagawa, T., Ohyama, Y., et al., 2008, PASJ, 60, 489

Joblin, C., Tielens, A.G.G.M., Allamandola, L.J., & Geballe, T.R., 1996, ApJ, 458, 610

Jones, T.J., Hyland, A.R., & Allen, D.A., 1983, MNRAS, 205, 187

Jones, A.P., 1990, MNRAS, 247, 305

Koike, C., Wickramasinghe, N.C., Kano, N., et al., 1995, MNRAS, 277, 986

Lee, W., & Wdowiak, T.J., 1993, ApJ, 417, L49

Mason, R.E., Wright, G., Pendleton, Y., & Adamson, A., 2004, ApJ, 613, 770

Mason, R.E., Geballe, T.R., Packham, C., et al., 2006, ApJ, 640, 612

McFadzean, A.D., Whittet, D.C.B., Bode, M.F., Adamson, A.J., & Longmore, A.J.,1989, MNRAS, 241, 873

Mennella, V., Brucato, J.R., Colangeli, L., & Palumbo, P., 1999, ApJ, 524, L71

Mennella, V., Brucato, J.R., Colangeli, L., & Palumbo, P., 2002, ApJ, 569, 531



Ott, U., 1993, Nature, 364, 25

Patzer, A.B.C., Chang, C., John, M., Sedlmayr, E., & Sulzle, D., 2003, Astron. Nachr.Suppl., 324, 20

Pendleton, Y.J., & Allamandola, L.J., 2002, ApJS, 138, 75

Pendleton, Y.J., Sandford, S.A., Allamandola, L.J., Tielens, A.G.G.M., & Sellgren, K.,1994, ApJ, 437, 683


Pino, T., et al., 2008, A&A, 490, 665

Pirali, O., Vervloet, M., Dahl, J.E., et al., 2007, ApJ, 661, 919

Risaliti, G., et al., 2006, MNRAS, 365, 303

Sakata, A., Wada, S., Onaka, T., & Tokunaga, A.T., 1990, ApJ, 353, 543

Sandford, S.A., Allamandola, L.J., Tielens, A.G.G.M., et al., 1991, ApJ, 371, 607

Sandford, S.A., Pendleton, Y.J., & Allamandola, L.J., 1995, ApJ, 440, 697

Schnaiter, M., Mutschke, H., Dorschner, J., Henning, T., & Salama, F., 1998, ApJ, 498,486

Scott, A.D., Duley, W.W., & Jahani, H.R., 1997, ApJ, 490, L175

Sellgren, K., Werner, M.W., & Ingalls, J.G., 2009, Amer. Astron. Soc. Meet. Abs., 214,#402.12

Sellgren, K., Werner, M.W., Ingalls, J.G., et al., 2010, ApJ, 722, L54

Sloan, G.C., et al., 2007, ApJ, 664, 1144

Szczerba, R., Siodmiak, N., & Szyszka, C., 2005, Planet. Nebulae Astron. Tools, 804, 214

Tielens, A.G.G.M., Wooden, D.H., Allamandola, L.J., Bregman, J., & Witteborn, F.C.,1996, ApJ, 461, 210


Volk, K., Kwok, S., Hrivnak, B., & Szczerba, R., 2001, Astrophys. Space Sci. Libr., 265,323


INTERACTION OF ATOMIC HYDROGEN WITH CARBONGRAINS

V. Mennella1

Abstract. Laboratory studies on the interaction of atomic hydrogenwith aliphatic and aromatic hydrogenated carbon grains are discussed.When exposed to atomic hydrogen, both types of hydrogenated carbongrains act as catalysts for molecular hydrogen formation. In the firstcase, an exchange reaction with hydrogen chemisorbed in aliphatic car-bon sites is the formation route to H2 formation. For aromatic carbongrains, the formation of molecular hydrogen takes place through a two-step reaction sequence: 1) super hydrogenation of the aromatic carbonislands of grains 2) exchange reactions on these islands. This mecha-nism represents a good approximation of molecular hydrogen formationon large neutral PAHs.

1 Introduction

The evolution of carbon dust in the Interstellar Medium (ISM) is driven by heat,irradiation by cosmic rays and UV photons and interaction with gas. Simulationof carbon dust processing under diffuse and dense medium conditions indicatesthat the key process for the evolution of the interstellar aliphatic carbon compo-nent is its interaction with H atoms. This interaction is able to counteract thedestruction of C-H bonds by UV photons and cosmic rays and activates the 3.4 μmabsorption band in diffuse interstellar regions and the 3.47 μm feature in denseclouds (Mennella et al. 2001, 2002, 2003; Mennella 2008a). The interesting aspectis that the same carbon grain population can absorb at the two different wave-lengths as a consequence of evolutionary transformations caused by processing.The transformations are compatible with the time-scale required by fast cyclingof materials between dense and diffuse regions of the ISM. Moreover, this carboncomponent, after the inclusion in a comet during the formation process in the coldouter edge of the solar nebula, may evolve to develop the CH2 and CH3 groups(as suggested by laboratory simulations) and contribute to the aliphatic band at

1 INAF-Osservatorio Astronomico di Capodimonte, Napoli, Italy



3.4 μm of IDPs and in particles of comet Wild 2 collected by the Stardust mission(Mennella 2010).

The interaction of gas with dust grains also plays an important role for thechemical evolution of the ISM. Many molecules observed in the ISM can form asa consequence of gas-phase reactions (Herbst et al. 2005 and references therein).However, the formation of important species (e.g. H2, H2O and CO2) requiresthe presence of dust grain surfaces that act as catalysts. In the case of molecularhydrogen, gas-phase reactions for the conversion of H atoms to H2 are inefficientto account for the observed abundances of this molecule in the ISM (e.g. Williams2005). The catalytic role of silicates, carbon grains and water ice has been demon-strated at low temperature through experiments performed by several groups (e.g.Pirronello et al. 1997, 1999; Manico et al. 2001; Hornekaer et al. 2003; Peretset al. 2005). The recombination efficiency is high only at low temperatures. Lab-oratory results fail to explain H2 production at higher grain temperatures suchthose expected in photodissociation regions (e.g. Habart et al. 2004), since theresidence time of H atoms in physisorption sites on the surface is too short forrecombination to take place.

2 H2 formation on aliphatic hydrogenated carbon grains

To explain H2 production at high grain temperatures, Cazaux & Tielens (2004)proposed H atom recombination in chemisorption sites of carbon particles. Theirmodel relies on the experimental and theoretical results obtained on aromaticstructures such as graphite and coronene-like carbon clusters. H atoms chemisorbedin aliphatic carbon sites were neglected, since their binding energy was consideredtoo high for those atoms to take part to molecular hydrogen formation. Recently,it has been shown experimentally that the formation of H2 in aliphatic sites ofhydrogenated carbon grains occurs (Mennella 2008b). Exposure of fully hydro-genated carbon grains with D atoms and deuterated carbon grains with H atomshas been performed. Figure 1 shows the spectral effects induced by atoms at300 K interacting with carbon grains at room temperature. The spectrum of hy-drogenated carbon grains (Fig. 1A) is characterized by the C-H stretching band at3.4 μm. Increasing the D atoms fluence, the C-H stretching feature decreases whilethe corresponding C-D feature increases. The analogous result is obtained whendeuterated carbon grains are exposed to H atoms (Fig. 1B). In this case, the C-Dfeature decreases while the C-H band increases when H atom fluence increases.

The reaction sequence governing the process is an exchange reaction: first,there is abstraction of hydrogen by a D atom with production of a HD moleculein gas phase; the second step is addition of D atom. HD molecules were de-tected with real-time mass spectroscopy. From the evolution with atom fluenceof the stretching band intensity the abstraction and addition cross sections havebeen estimated. The small value of the abstraction cross section (3×10−18 cm2)indicates that the formation of HD takes place through an Eley-Rideal process(see Mennella 2008b for more details). Moreover, abstraction is the rate limitingstep of the reaction sequence, since the abstraction and addition cross sections

V. Mennella: H Atom Interaction with Carbon Grains 395

Fig. 1. Evolution of the C-H and C-D stretching bands during D irradiation of hydro-

genated carbon grains (A) and H irradiation of deuterated carbon grains (B). In both

panels the initial spectrum (a) and those after irradiation (shown after subtraction of the

initial spectrum) of 6 × 1017 (b) and 1 × 1019 atoms cm−2 (c) are plotted. The differ-

ence spectra are offset for clarity. The weak sp3 CH2,3 band present in the spectrum of

deuterated carbon grains is due to the small fraction (2%) of H2 present in the D2 gas

used to produce the grains.

are equal, within the errors. These experimental results clearly suggest that H2

formation occurs on carbon grains even at high grain temperatures since H atomsare bonded to carbon with an energy of about 4 eV, and do not desorb even athigh temperatures.

3 H2 formation on aromatic hydrogenated carbon grains

Theoretical studies have suggested that H2 can form from abstraction of edgeH atoms of PAH cations by incoming H atoms (Cassam-Chenai et al. 1994).Alternatively, it has been proposed that H2 can form on PAH cations in a two-step process: 1) an extra H atom is added to an edge site 2) abstraction of theexcess H atom by a second incoming H atom (Bauschlicher 1998). In the caseof cations, the formation does not have significant energy barriers. Theoreticalstudies have shown that low barrier routes to molecular hydrogen formation onneutral coronene exist: H2 should form by H abstraction on super hydrogenatedcoronene (Rauls & Hornekaer 2008; Thrower et al., this volume). Laboratoryexperiments have shown that atomic and molecular hydrogen react with benzeneand small PAHs (Petrie et al. 1992; Scott et al. 1997; Le Page et al. 1997; Snowet al. 1998). However, the formation of H2 has not been observed as yet.


Fig. 2. IR spectrum of BE carbon grains as produced (a). The spectrum is characterized

by the C-C and C-O bands at 1600 cm−1 (6.25 μm) and 1710 cm−1 (5.85 μm) respectively,

the weak band at 1442 cm−1 (6.93 μm), and by the aromatic C-H bands at 3050 cm−1

(3.28 μm) (stretch), the weak mode at 1157 cm−1 (8.64 μm) (in plane bend) and the out

of plane bend modes between 930 and 670 cm−1. The spectrum after irradiation (shown

after subtraction of the initial spectrum) of 1.65×1019 D atoms cm−2 (b) is also plotted.

The intensity of the C-C and C-H bands decreases while the aliphatic C-D stretch bands,

falling between 2280 and 2000 cm−1, and the corresponding bending modes, between

1120 and 1020 cm−1, develop with D exposure.

Recently, the interaction of H atoms with a carbon soot (BE) with a markedaromatic character has been studied experimentally (Mennella, in preparation).BE grains have been produced by burning benzene (C6H6) in air; they are char-acterized by a size distribution of ringed sp2 islands (e.g. Colangeli et al. 1995).The infrared spectrum of BE (Fig. 2) indicates that, with the exception of thefeatures at 6.25 μm and 5.85 μm, all the bands are due to C-H vibrations of hy-drogen bonded to aromatic carbon sites. There is no evidence for aliphatic C-Hvibrational modes.

Exposure of BE at room temperature to D atoms at 300 K induces signifi-cant changes in the spectrum (Fig. 2). The C-C feature at 6.25 μm and all thebands due to the aromatic C-H modes decrease in intensity with D fluence. Theformation of the C-D aromatic bands is not observed, while the aliphatic C-Dbands develop with D irradiation. The feature at 2912 cm−1 is attributed to theC-H stretching mode of CHD. The detailed evolution with D atom fluence of thearomatic C-H and aliphatic C-D stretching band intensity is shown in Figure 3.From the observed trends, it is possible to derive the destruction cross section ofaromatic C-H bonds and the formation of the aliphatic C-D bonds. The small

V. Mennella: H Atom Interaction with Carbon Grains 397

Fig. 3. Evolution with D atom fluence of the aromatic C-H (filled circles) and aliphatic

C-D (empty cicles) stretch band integrated optical depth. The best fit to the data

of the relations a1[1 − exp(−σfastF )] + a2[1 − exp(−σslowF )] and b1 exp(−σfastF ) +

b2 exp(−σslowF ), respectively for the C-D and C-H modes, are also shown. The estimated

cross sections are expressed in units of 10−2 A2.

values of the cross sections suggest that an Eley-Rideal type process drives thespectral changes produced by D atom irradiation. Unlike the case of D irradia-tion of aliphatic hydrogenated carbon grains, the spectral changes do not suggestan exchange reaction in the aromatic sites, since the formation of aromatic C-Dbonds is not observed. Rather, they indicate super hydrogenation of the aromaticstructures of grains as new aliphatic C-D bonds form. In fact, D atoms break asp2 double bonds inducing the formation sp3 C-D bonds.

As one can see in Figure 3, at high fluences the spectral variations are smallerand smaller. To understand which reactions are active in those conditions, at theend of D atom exposure of BE, irradiation has been continued using H atoms at300 K. H atom exposure determines an increase of the aliphatic C-H stretchingand bending modes and a decrease of the corresponding aliphatic C-D bands.On the contrary, there is no variation of the aromatic features. These spectralchanges are similar to those observed during H atom irradiation of deuteratedaliphatic carbon grains and are indicative of exchange reactions in the aliphaticsites. The exchange reactions form HD molecules. These results clearly show thatsuperhydrogenated aromatic carbon islands act as catalysts for molecular hydrogenformation. Moreover, they represent a good approximation of H2 formation onneutral PAHs, as the reactions induced by H atoms in aromatic carbon grains areEley-Reideal type processes.

This work has been supported by ASI research contracts.


References

Bauschlicher, C.W., Jr., 1998, ApJ, 509, L125

Cassam-Chenaı, P., Pauzat, F., & Ellinger, Y., 1994, Molec. Grains Space, 312, 543

Cazaux, S., & Tielens, A.G.G.M., 2004, ApJ, 604, 222

Colangeli, L., Mennella, V., Palumbo, P., Rotundi, A., & Bussoletti, E., 1995, A&AS,113, 561

Habart, E., Boulanger, F., Verstraete, L., Walmsley, C.M., & Pineau des Forets, G.,2004, A&A, 414, 531

Herbst, E., Chang, Q., & Cuppen, H.M., 2005, J. Phys. Conf. Ser., 6, 18

Hornekær, L., Baurichter, A., Petrunin, V.V., Field, D., & Luntz, A.C., 2003, Science,302, 1943

Le Page, V., Keheyan, Y., Bierbaum, V.M., & Snow, T.P., 1997, J. Am. Chem. Soc.,119, 8373

Manico, G., Ragunı, G., Pirronello, V., Roser, J.E., & Vidali, G., 2001, ApJ, 548, L253

Mennella, V., 2010, ApJ, 718, 867

Mennella, V., 2008a, ApJ, 682, L101

Mennella, V., 2008b, ApJ, 684, L25

Mennella, V., Baratta, G.A., Esposito, A., Ferini, G., & Pendleton, Y.J., 2003, ApJ, 587,727


Mennella, V., Munoz Caro, G., Ruiterkam, R., et al., 2001, A&A, 367, 355

Perets, H.B., Biham, O., Manico, G., et al., 2005, ApJ, 627, 850

Petrie, S., Javahery, G., & Bohme, D.K., 1992, J. Am. Chem. Soc., 114, 9205

Pirronello, V., Biham, O., Liu, C., Shen, L., & Vidali, G., 1997, ApJ, 483, L131

Pirronello, V., Liu, C., Roser, J.E., & Vidali G., 1999, A&A, 344, 681

Rauls, E., & Hornekær, L., 2008, ApJ, 679, 531

Scott, G.B., Fairley, D.A., Freeman, C.G., et al., 1997, J. Phys. Chem. A, 101, 4973


Williams, D.A., 2005, J. Phys. Conf. Ser., 6, 1


NEAR-INFRARED SPECTROSCOPY OF INTERSTELLARDUST

F. Boulanger1, T. Onaka2, P. Pilleri3, 4 and C. Joblin3, 4

Abstract. Near infrared observations of reflection nebulae have set thehistorical ground for the discovery of interstellar PAHs, but since, spaceobservations have focused on their mid-IR features, and data shortwardof 5μm have remained scarce. The Spitzer/IRAC images in the 3.6 and4.5 μm channels do show that the near-IR emission from small dustparticles is ubiquitous across the Galaxy, but provide no spectroscopicinformation. To investigate the nature of this near-IR dust emission, wehave obtained AKARI spectroscopic observations, over the 2.5 − 5 μmspectral range, for a set of archetype PDRs mapped with the Spitzerspectrometer at mid-IR wavelengths. These AKARI data supplementearlier observations with the SWS ISO spectrometer, in providing thegain in sensitivity needed to observe low excitation sources, and thespatial information required to spatially correlate near-IR spectroscopicsignatures with physical conditions and observed changes in mid-IRspectra. This paper presents the first results of the data analysis, inrelation to two open questions on interstellar PAHs. (1) Is there anevolutionary link from aliphatic carbon dust to PAHs? (2) What isthe origin of the near-IR dust continuum? The AKARI spectra displayfeatures longward of the main 3.29 μm PAH feature, and continuumemission. The intensity ratio between the features ascribed to aliphaticCH bonds and the 3.29 μm aromatic band, varies spatially in a way thatmay be interpreted as evidence for aromatization of the smallest dustparticles by photo-processing. The continuum displays a striking step-increase across the 3.29 μm feature. We also present a spectrum ofa photodissociation region with a feature at 4.65 μm, which has beenspeculated to be related to the CD stretch in aliphatic hydrocarbonside-groups on PAHs.

1 Institut d’Astrophysique Spatiale (IAS), UMR 8617, CNRS & Universite Paris-Sud 11,Batiment 121, 91405 Orsay Cedex, France2 Department of Astronomy, University of Tokyo, Tokyo 113-0033, Japan3 Universite de Toulouse, UPS, CESR, 9 avenue du colonel Roche, 31028 Toulouse, France4 CNRS, UMR 5187, 31028 Toulouse, France



1 Introduction

The discovery of near infrared (Near-IR) extended emission in reflection nebulae(Sellgren et al. 1983) initiated a new field of observations and research on inter-stellar dust. To account for the 3.3 μm feature and the associated mid-IR featuresLeger & Puget (1984) proposed that the emission follows stochastic heating ofpolycyclic aromatic hydrocarbon molecules (PAHs). The PAH emission featureshave since been observed to be a ubiquitous spectroscopic signature of dust ingalaxies. Very small carbon dust particles (PAH molecules and very small grains)are established as a major constituent of interstellar dust that play key roles inthe physics and chemistry of the interstellar medium (ISM), but discussions atthis conference have shown that the origin and nature of these particles are stilldebated.

Near-IR wavelengths are of particular relevance to characterize the smallestparticles of interstellar dust. First, the near-IR dust emission is thought to arisefrom the very smallest particles with less than ∼100 carbon atoms. Second, emis-sion in the C-H stretching modes in the 3.3 to 3.5 μm range are best suited toprobe the presence of aliphatic side groups to PAHs. Other characteristic modesof CH3 and CH2 bonds that fall at ∼6.9 and ∼7.2 μm are more difficult to observedue to the presence of the 6.2 and 7.7 μm PAH emission band. Third, one mayprobe the physics of the particle emission by characterizing the emission shortwardand longward of the C-H stretch bands.

Despite its specific interest, the near-IR spectral range has been barely explored(Geballe et al. 1989). The IRAS, IRTS, ISO and Spitzer space missions haveprovided us with a wealth of mid-IR observations, but spectroscopy shortward of5 μm is only available for bright sources with a high UV field (Verstraete et al.2001; van Diedenhoven et al. 2004), such as the Orion Bar and the M17SW photo-dissociation regions (PDRs), and with little spatial information. The AKARIwarm mission has provided the opportunity to obtain near-IR spectroscopy ofdiffuse emission with unprecedented sensitivity (Ohyama et al. 2007). In thispaper, we briefly present our AKARI observing program, and the first results ofthe data analysis. This is an early account of on-going work, which will lead to amore thorough publication in a refereed journal.

2 AKARI observations

Our AKARI observing program was designed to investigate dust emission in PDRs.We used the AKARI grating spectrometer with an entrance slit 50′′ long. Ourtargets comprise 15 slit positions selected within six archetype star forming re-gions, on the basis of ISO and Spitzer mid-IR spectro-imaging observations. Thelist of targets includes objects with different excitation conditions (UV radiationfield from 100 to a few 1000 times the Solar Neighborhood value), which showstrong variation in their mid-IR spectra (C-C to C-H features intensity ratio andband/continuum contrast). It also spans a range of distances, from the local ISMto the 30 Doradus LMC star forming region, to explore a range of physical scales.

F. Boulanger et al.: Near-Infrared Spectroscopy of Interstellar Dust 401

Fig. 1. AKARI observation of one pointing towards the reflection nebula NGC 7023.

The spectral image is displayed to the left (wavelengths increasing from the left to the

right). To the right, we show the spectrum extracted by averaging data within the two

red lines at the top of the image. The log-scale highlights the fainter features at longer

wavelengths, and the near-IR continuum with its step-increase across the 3.3 − 3.5 μm

range.

These AKARI data offer the combination of spatial (along the slit) and spectralinformation, required to spatially correlate near-IR spectroscopic signatures withphysical conditions and mid-IR spectro-imaging observations.

The data have been reduced and calibrated by the Japanese team membersusing the best knowledge of the instrument (see Onaka et al. elsewhere in this vol-ume). The outcome of the data reduction is a spectral image from which spectracan be extracted as a function of position along the slit (Fig. 1). The spectral reso-lution, measured on hydrogen recombination lines, ranges from 90 to 160. We usean observation of HII gas emission towards 30 Doradus to validate the spectrom-eter capability to perform photometry of extended emission. This HII spectrumshows 10 hydrogen lines and continuum emission (free-free + free-bound), whichare both in agreement with predictions from photo-ionization models. Spectrallines with intensities down to 2–3 MJy/sr, and a 5 MJy/sr continuum, are reliablymeasured.

3 AKARI dust spectra

To illustrate the initial results of the data analysis, we present spectra measured forone slit position in the reflection nebula NGC 7023 (Figs. 1 and 2), and a spectrumof a molecular cloud in the Large Magellanic Cloud (Fig. 3). This last spectrumobtained towards the massive star forming region 30 Doradus includes severalhydrogen recombination lines. In Figure 3, the gas emission, lines and continuum,has been removed using a photo-ionization model. The near-IR emission fromdust comprises the prominent 3.29 μm feature, a weaker spectrally structuredplateau extending to 3.6 μm, and continuum emission. The step increase of thecontinuum emission from short to long wavelengths, across the 3.3−3.5 μm range,is a remarkable result, which is observed in all spectra.

The 3.29 μm feature is assigned to the v = 1 – 0 aromatic C–H stretch tran-sition. Emission in this feature is predicted to come mainly from neutral PAHs(DeFrees et al. 1993). Several types of transitions, most likely the C–H stretch


Fig. 2. AKARI spectra at two positions within the reflection nebula NGC 7023. The

two positions sample the variation in intensity ratios of the mid-IR PAH features, which

has been interpreted as evidence of the transition from neutral to cation PAHs (Rapacioli

et al. 2005; Berne et al. 2007). From the left to the right spectra, the AKARI observa-

tions show the disappearance of the 3.4 μm emission feature, and an increase by a factor

2 of the intensity of the continuum relative to that of the 3.3 μm emission feature.

in aliphatic side groups attached to PAHs or in sur-hydrogenated PAHs, but alsopossibly, combination bands and hot bands including the v = 2 – 1 aromatic C–Htransition, also contribute to the plateau and superimposed features (Allamandolaet al. 1989; Joblin et al. 1996; Bernstein et al. 1996). The near-IR continuumhas been observed to be a dominant component of the near-IR emission from thediffuse ISM (Flagey et al. 2006), but it has never received a definitive interpre-tation within the PAH hypothesis. The step-increase in the continuum intensityacross the 3.29 μm (Fig. 1) band strongly suggests that the continuum is asso-ciated with the de-excitation of PAHs, but the emission process has still to becharacterized. Allamandola et al. (1989) proposed that the continuum repre-sents emission though vibrational modes, which become weakly active in excitedPAHs. The combination of a large number of weak overlapping bands would makea quasi-continuum with little spectral structure. Within this interpretation, thestep increase of the continuum emission across the C-H 3.29 μm feature could re-flect the lower coupling efficiency between C-C and C-H modes than between C-Cmodes. Fluorescence emission from an electronic excited state provides an alter-native interpretation (Leger et al. 1988). Modelling (private communication fromPino and Schmidt) supported by experimental data suggests that, in PAH cations,the energy gap between the first excited state and the ground state may be smallenough for cations to fluoresce in the near-IR. A more detailed analysis of the datawill allow us to test this interpretation by looking for a spatial correlation betweenthe continuum emission and that of PAH cations inferred from the decompositionof mid-IR spectra (Rapacioli et al. 2005; Berne et al. 2007).

4 Evidence for small dust processing

The nature of the small carbon dust particles has to be addressed within thebroad context of the evolutionary cycle that can link them to the amorphous


Fig. 3. AKARI spectrum of a molecular cloud in the Large Magellanic Cloud. The

plot shows the emission spectrum with the dust and hydrogen lines to the left, and the

spectrum obtained after subtraction of the gas emission to the right. We are confident

that the 4.65 μm feature is real. We have no interpretation for this band at present.

hydrogenated carbon dust seen in absorption spectroscopy (Dartois in this vol-ume). Interstellar PAHs and VSGs could be produced by photo and/or mechanicalfragmentation in grain-grain collisions involving amorphous carbon grains (Joneset al. 1996) or PAH clusters (Rapacioli et al. 2006). The free fragments are ex-pected to be progressively photo-processed by selective dissociation of side groupsand loss of non-aromatic CH bonds, when they are exposed to increasing ultravi-olet photon fluxes as they move out of the shielded interiors of molecular clouds.This evolutionary scenario can be tested by studying the evolution of the 3.4 μmaliphatic band relative to the 3.3 μm aromatic band. In particular, we plan tostudy whether these variations are spatially correlated with variations in emissioncomponents inferred from mid-IR spectro-imaging observations (Rapacioli et al.2005; Berne et al. 2007).

Onaka et al. (elsewhere in this volume) report large variations of the emissionratio, 3.4+3.5 μm to 3.3 μm, for the diffuse Galactic emission. We observe similarvariations on smaller scales within nearby PDRs. These results extend ground-based observations on brighter sources (Geballe et al. 1989; Joblin et al. 1996).The two spectra of NGC 7023 in Figure 2 are taken at the opposite ends of the 50′′

long slit, within and out of the dense surface of the molecular cloud (left and rightspectra, respectively), marked by bright filamentary structure in PAH and near-IRH2 images. The disappearance of the 3.4 μm emission feature may be interpretedas evidence for aromatization of the smallest dust particles by photo-processing.A thorough analysis of the spatial variations of the aliphatic to aromatic emissionratio across the NGC 7023 PDR is required to ascertain and fully quantify thisinterpretation.

5 Looking for new spectral features

The AKARI spectra offer an unprecedented opportunity to look for faint near-IRspectral features. The C–D stretching mode around 4.5 μm (4.4 and 4.65 μm for


the aromatic/aliphatic C–D bond, respectively) is of particular interest to test ifPAHs are highly deuterated (Draine 2004; Peeters et al. 2004). The C–D in-plane and out-of-plane bending modes, expected at 11.7 and 15.4 μm, cannot beunambiguously identified, because they fall in a spectral region where other PAHfeatures are present.

Analyses of spectra obtained with the Far Ultraviolet Spectroscopic Explorer(FUSE) satellite, together with spectra from the Copernicus space mission andISM absorption probe spectrograph (IMAPS), have revealed a factor 4 variationin the observed D/H ratios for interstellar gas in the Solar Neighborhood (Linskyet al. 2006). Linsky et al. argue that spatial variations in the depletion ofdeuterium onto dust grains can explain the variations in the observed gas-phaseD/H ratios. This interpretation is supported by the correlation of D/H with the gascolumn density and depletions of the refractory metals iron and silicon. Gas-grainreactions could lead to a high deuteration of hydrogenated carbonaceous grainsand PAHs (Draine 2004). The gain in sensitivity provided by AKARI permit tolook for deuterated PAHs in sources, less excited than the Orion Bar and M17SWPDRs studied in Peeters et al. (2004), which are more directly relevant to thedeuteration of PAHs in the diffuse ISM (Onaka et al. in this volume).

We are presently cautious in reporting new spectral features, but we are notaware of any data artifact that could contaminate the emission feature at 4.65 μmin Figure 3. We are thus confident that this feature is real. Its central wavelengthmatches that of the aliphatic C–D stretch mode (Peeters et al. 2004). However,the absence of the corresponding aromatic band at 4.4 μm, and the intensity ofthe feature – comparable to that of the 3.4 μm – make the identification withdeuterated aliphatic side groups on PAHs unsure.

We thank several participants of the conference, in particular Lou Allamandola, PhilippeBrechignac, Thomas Pino, Tim Schmidt, Kris Sellgren, and Xander Tielens, for discussionsthat helped us with the interpretation of the data.

References

Allamandola, L.J., Tielens, A., & Barker, J.R., 1989, ApJS, 71, 733

Berne O., Joblin, C., Deville, Y., et al., 2007, A&A, 469, 575

Bernstein, M.P., Sandford, S.A., & Allamandola, L.J., 1996, ApJ, 472, L127

DeFrees, D.J., Miller, M.D., Talbi, D., et al., 1993, ApJ, 408, 530

Draine, B.T., 2004, in Astrophysics in the Far Ultraviolet [Astro-ph/0410310]

Draine, B.T., & Li, A., 2007, ApJ, 657, 810

Flagey N., Boulanger F., Verstraete, L., et al., 2006, A&A, 453, 969

Geballe, T.R., Tielens, A.G.G.M., Allamandola, L.J., et al., 1989, ApJ, 341, 278

Joblin, C., Tielens, A.G.G.M., Allamandola, L.J., & Geballe, T.R., 1996, ApJ, 458, 610

Jones, A.P., Tielens, A.G.G.M., & Hollenbach, D.J., 1996, ApJ, 469, 740


Leger, dHendecourt, L., & Boissel, P., 1988, PhRvL, 60, 921L


Linsky, J., Draine, B.T., Moos, H.W., et al., 2006, ApJ, 647, 1106

Ohyama, Y., Onaka, T., Matsuhara, H., et al., 2007, PASJ, 59, S411

Peeters, E., Allamandola, L.J., Bauschlicher, C.W., et al., 2004, ApJ, 604, 252


Rapacioli, M., Calvo, F., Joblin, C., et al., 2006, A&A, 460, 519

Sellgren, K., Werner, M.W., & Dinerstein, H.L., 1983, ApJ, 271, L13


Verstraete, L., Pech, C., Moutou, C., et al., 2001, A&A, 372, 981


ATYPICAL DUST SPECIES IN THE EJECTA OF CLASSICALNOVAE

L.A. Helton1, A. Evans2, C.E. Woodward1 and R.D. Gehrz1

Abstract. A classical nova outburst arises from a thermonuclear run-away in the hydrogen-rich material accreted onto the surface of awhite dwarf in a binary system. These explosions can produce copiousamounts of heavy element enriched material that are ejected violentlyinto the surrounding interstellar medium. In some novae, conditions inthe ejecta are suitable for the formation of dust of various compositions,including silicates, amorphous carbon, silicon carbide, and hydrocar-bons. Multiple dust grain types are sometimes produced in the samesystem. CO formation in novae may not reach saturation, thus in-validating the usual paradigm in which the C:O ratio determines thedust species. A few novae, such as V705 Cas and DZ Cru, have ex-hibited emission features near 6, 8, and 11 μm that are similar to“Unidentified Infrared” (UIR) features, but with significant differencesin position and band structure. Here, we present Spitzer IRS spectraof two recent dusty novae, V2361 Cyg and V2362 Cyg, that harborsimilar peculiar emission structures superimposed on features arisingfrom carbonaceous grains. In other astronomical objects, such as starforming regions and young stellar objects, emission peaks at 6.2, 7.7,and 11.3 μm have been associated with polycyclic aromatic hydro-carbon (PAH) complexes. We suggest that hydrogenated amorphouscarbon (HAC) may be the source of these features in novae based uponthe spectral behavior of the emission features and the conditions underwhich the dust formed.

1 Introduction

A classical nova (CN) is an eruption arising on the surface of a white dwarf (WD)in a binary system in which the WD is accreting material from a main sequence

1 Department of Astronomy, School of Physics and Astronomy, University of Minnesota, 116Church St. SE, Minneapolis, MN 55455, USA;e-mail: [email protected]; [email protected]; & [email protected] Astrophysics Group, Keele University, Keele, Staffordshire ST5 5BG, UK;e-mail: [email protected]



companion through Roche-lobe overflow. The accreted material accumulates onthe surface of the WD until it reaches a density and temperature high enoughto initiate a thermonuclear runaway (TNR), resulting in explosive nucleosynthesisand violent ejection of up to 10−6 to 10−4 M� of accreted material at velocitiesexceeding 1000 km s−1. The ejecta are initially optically thick, but as they ex-pand, the density rapidly declines resulting in the recession of the optically thickpseudophotosphere inwards towards the hot underlying WD. This energetic under-lying central engine actively photoionizes the ejecta, resulting in the appearanceof emission lines from a wide range of isoelectronic species.

CNe frequently produce dust in their ejecta. The dust condensation event issignaled by a rapid decline in the optical light curve due to extinction and a cor-responding rise in the infrared luminosity due to thermal emission. Models of thespectral energy distribution (SED) in these systems suggest that the condensationevent yields relatively large dust grains with radii up to ∼1 μm (Gehrz 2008).Unlike most astronomical sources of dust, novae have been observed to produceboth carbon rich and oxygen rich dust species simultaneously (e.g., Gehrz et al.1992). Evans & Rawlings (2008) suggest that this chemical dichotomy may beexplained by incomplete carbon monoxide formation in the ejecta. If CO does notform to saturation, then neither carbon nor oxygen will be entirely bound up inCO, leaving both available for dust production.

2 Atypical dust species in classical novae

Novae V705 Cas, V842 Cen, and QV Vul exhibited emission features similar tounidentified infrared features (UIRs; Geballe 1997) often attributed to polycyclicaromatic hydrocarbons (PAHs; Allamandola et al. 1989) or hydrogenated amor-phous carbons (HACs; Duley & Williams 1983). Observations of V705 Cas be-ginning 157 days after outburst revealed UIR features at 8.2, 8.7, and 11.4 μmsuperimposed on a 10 μm silicate emission feature (Evans et al. 1997). Subse-quent analysis of the 8 μm complex by Evans et al. (2005) indicated that the bandshape and peak wavelength did not match those of standard UIR features and thatthe peak wavelength drifted bluewards as the system evolved. Simultaneously, theratio of the 11.4 to 8 μm features increased. The 3–4 μm region in the spectrumof V705 Cas showed emission features at 3.28 and 3.4 μm. These are typicallyattributed to unsaturated C-H stretching mode oscillations in aromatic moleculesand saturated C-H stretching vibrations in aliphatic bonds, respectively. Thus,the 3.28/3.4 μm ratio can be used as an indicator of the relative abundance ofaromatic to aliphatic structures in the emitting material (Sloan et al. 1997). Theobserved 3.28 to 3.4 μm ratio was similar to that seen in post-AGB stars, but anorder of magnitude higher than in stars with a high UV photon flux. The behav-ior of these features in V705 Cas was similar to those observed in V842 Cen andQV Vul. Evans & Rawlings (1994) suggested that during the optically thick stageof dust formation, the rate of H capture is greater than the rate of UV photonabsorption. Thus, the conditions are conducive to the growth of aliphatic bonds,e.g. HAC grains.

L.A. Helton et al.: Atypical Dust in the Ejecta of Novae 409

Fig. 1. Left: Spitzer IRS spectra of DZ Cru with a DUSTY model of AC at 470 K.

Right: residuals after DUSTY model subtraction showing UIR features.

2.1 DZ Cru

DZ Cru was observed as part of our Spitzer Classical Novae Target-of-Opportunityand Monitoring campaign (Evans et al. 2010). Due to poor sampling of the lightcurve, the optical depth of the dust extinction event is not known. Our first SpitzerInfrared Spectrograph (IRS) observations commenced nearly 1500 days after erup-tion. They revealed exceptionally strong emission from amorphous carbon (AC)with superimposed UIR emission features reminiscent of those observed in V705Cas. The left panel of Figure 1 shows three epochs of Spitzer IRS data along witha DUSTY (Ivezic et al. 1999) model of AC dust at 470 K. Subtraction of theDUSTY model (Fig. 1, right panel) revealed emission features at 6.5, 7.2, 8.1, 9.3,11.0, and 12.4 μm. The wavelengths of these features do not correspond with thestandard PAH features observed at 6.2, 7.7, 8.6, 11.3, and 12.7 μm.

2.2 V2362 Cyg

V2362 Cyg was a peculiar nova in many respects (see Lynch et al. 2008). Oneof the most interesting characteristics was its optical light curve behavior. Aftera smooth initial decline the light curve plateaued and subsequently experienceda secondary brightening event, which rivaled its initial outburst luminosity. Atthe peak of the second outburst, the nova underwent a rapid and extreme dustformation event. The Spitzer IRS spectra obtained immediately following dustcondensation revealed blackbody emission at 1400 K, near the extreme high endof dust condensation temperatures. At first, low ionization [Ne II], [Ne III], and[O IV] were observed in the spectra alongside a hydrogen recombination spectrum.But within a few days, as the dust emission strengthened, the emission lines dis-appeared with the sole exception of [O IV], which remained strong throughout.

With the decline in the dust optical depth due to expansion of the ejecta,the emission lines reappeared, but at a much higher ionization level. A range ofionization states was observed including [Ne III], [Ne V], [Ne VI], [Mg V], and[Mg VII], indicating a very energetic ionization field. Indeed, Swift observations


from this period indicated a strong, hard X-ray flux that persisted for at least300 days.

During this same period of optically thin dust emission, strong, broad UIRfeatures near 6–10, 11.5, and 18 μm became evident. We fitted and subtracted ablackbody continuum to isolate the emission components. The results are shownin the left panels of Figure 2. Distinct emission peaks appeared at 6.4, 6.9, 7.9,9.5, 11.4, 12.5, and 17.9 μm. The features near 9.5 and 18 μm may be associatedwith silicates. As the ejecta evolved, the central wavelengths of the UIR featuresvaried slightly and irregularly, while the flux ratios changed dramatically. The6.9, 11.4, and 12.5 μm features declined rapidly and soon disappeared. The fluxesof the 6.4 and 7.9 μm features declined sharply relative to the 9.5 and 17.9 μmfeatures. The 17.9 μm feature began to dominate the spectrum at late times.

2.3 V2361 Cyg

The spectrum of V2361 Cyg was very similar to that of V2362 Cyg. The primarydifferences were the lack of emission features at 9.5 and 18 μm and additionalstructure in the 6–10 μm spectral region. Spectra obtained as part of our Spitzerprogram were fitted with blackbody curves. The residuals after blackbody sub-traction are shown in the right panels of Figure 2. UIR emission features wereobserved at 5.4, 6.4, 6.9, 7.1, 7.5, 8.1, 11.5, and 12.4 μm. As the ejecta evolved,the complex of features between 6.4 and 8.1 μm blended together to form a singleplateau, while the 6.4 to 8.1 μm flux ratio diminished significantly. The 11.5 and12.4 μm features joined to produce a broad structure around 13.1 μm.

In Figure 2-h, we compare the observed emission from V2361 Cyg at 102 dayspost outburst to the Class A, B, and C PAH features described by Peeters et al.(2002) and van Diedenhoven et al. (2004). The Class A and B PAHs may beexcluded as potential contributors due to the poor fit to the emission peaks. TheClass C profiles are more consistent with the data. In particular, the 8 μm com-ponent of the Class C emission fits the red side of the profile quite well. Sloanet al. (2007) have attributed the Class C profiles to carriers with a high fractionof aliphatic bonds.

3 The model

The Spitzer IRS observations of DZ Cru, V2362 Cyg, and V2361 Cyg in combi-nation with the previous studies of V705 Cas, V842 Cen, and QV Vul provide thebasis for an elementary model for the processing of dust and complex hydrocarbonmolecules in classical novae. The dust formation event proceeds quite rapidly withdust grains condensing out of the gas phase and growing to sizes up to ∼1 μmin a matter of days. In part, this is likely due to the inhomogeneous, clumpydistribution of the ejecta. With the onset of grain formation, the opacity in theseregions increases, shielding the rest of the material from the underlying radiationand allowing the acceleration of condensation.

L.A. Helton et al.: Atypical Dust in the Ejecta of Novae 411

Fig. 2. Residuals for V2362 Cyg, (a)–(d), and V2361 Cyg, (e)–(g). Dashed lines are

individual Gaussian components of the composite fit shown in red. Panel (h) shows

Class A, B, and C PAH profiles (Peeters et al. 2002) overlayed on the V2361 Cyg spectra

from Day 102.

The types of dust produced during the condensation event depend on the condi-tions and abundances in the ejecta. In typical novae, the dust formed is composedof AC, silicates, or some combination thereof. In the novae discussed here, thedominant dust species appears to be carbonaceous. Evans & Rawlings (1994) ar-gued that at the opacities observed during the dust condensation stage in novae,the likelihood for H capture is much higher than the probability of UV photoninteraction, thus enhancing the production of aliphatic bonds. This conclusionsuggests that the smaller end of the grain size distribution may be due to PAHmolecules and larger, more complex HAC agglomerations. However, as the ejectaexpand and densities decline, the optical depth of the dust forming regions de-creases as the pseudophotosphere recedes, exposing the freshly condensed dust toan increasingly hard and intense radiation field.


Evans & Rawlings (1994) have calculated that small PAHs (i.e., molecules withfewer than 24 C atoms) have a very short lifetime in the environment of a CN.Due to the high UV flux, a free-flying PAH molecule would absorb about onehigh energy photon every second and would be completely disrupted in less thana day. These same molecules would also be subjected to chemical erosion throughcollisional interactions in the ejecta. For reasonable densities, temperatures, andabundances in the ejecta, Evans & Rawlings estimated that these PAHs would bedestroyed due to chemi-sputtering by H and O in a matter of hours.

If some of the emission features seen in V2361 Cyg, V2362 Cyg, and DZ Crucan be attributed to small PAHs instead of HACs, then there must be a reservoirof material from which the PAHs are being repopulated. This material could beproduced by either HACs or the larger AC grains themselves when submitted toenergetic processes. Though rapidly destroyed, the transient population of free-flying PAHs could generate some of the observed UIR emission features.

In order to better constrain the carriers of the UIR features observed in CN, itis critical to obtain observations of the expanding ejecta throughout the dust con-densation and subsequent stages. If the carrier molecules are being formed duringthe initial condensation event, they may be observable during or immediately aftercondensation. Data covering the 3–4 μm region are required in order to determinethe relative contributions of aromatic to aliphatic material.

This work was supported in part by Spitzer/JPL grants 1294825 and 1289430.

References


Duley, W.W., & Williams, D.A., 1983, MNRAS, 205, 67P

Evans, A., & Rawlings, J.M.C., 1994, MNRAS, 269, 427

Evans, A., Geballe, T.R., Rawlings, J.M.C., Eyres, S.P.S., & Davies, J.K., 1997, MNRAS,292, 192

Evans, A., Tyne, V.H., Smith, O., et al., 2005, MNRAS, 360, 1483

Evans, A., & Rawlings, J.M.C., 2008, in Classical Novae, 2nd Edition, ed. M.F. Bode &A. Evans (Cambridge Univ. Press: Cambridge)

Evans, A., Gehrz, R.D., Woodward, C.E., et al., 2010, MNRAS, 406, 85

Geballe, T.R., 1997, From Stardust to Planetesimals, 122, 119

Gehrz, R.D., 2008, in Classical Novae, 2nd Edition, ed. M.F. Bode & A. Evans(Cambridge Univ. Press: Cambridge)

Gehrz, R.D., 1992, ApJ, 400, 671

Ivezic, Z., Nenkova, M., & Elitzer, M., 1999 [arXiv:astro-ph/9910475]

Lynch, D.K., Woodward, C.E., Gehrz, R.D., et al., 2008, AJ, 136, 1815



Sloan, G.C., Bregman, J.D., Geballe, T.R., Allamandola, L.J., & Woodward, C.E., 1997,ApJ, 474, 735


The Role of PAHs

in the Interstellar Medium


THE ROLE OF PAHS IN THE PHYSICS OF THEINTERSTELLAR MEDIUM

L. Verstraete1

Abstract. In the interstellar medium (ISM), PAHs are abundant andalso carry most of the dust surface. They are thus privileged sites forsurface reactions such as the formation of H2. In regions penetratedby UV photons, PAHs loose electrons by the photoelectric effect andefficiently heat the gas. In more shielded regions, PAH recombine andmay carry an important fraction of the cloud electronic charge whichplays an important role in the gas dynamics and chemistry. We reviewhere processes involving PAHs which control key aspects of the physicsof the ISM. We also discuss the corresponding observational constraints.Most of these processes involve a detailed knowledge of the charge ofPAHs and we therefore review current models in this area. We arguethat more laboratory measurements of the rate of electronic captureon large PAH cations are needed.

1 Introduction

The interstellar lifecycle brings matter from low-density medium (the diffuse ISMor atomic HI gas heated by the standard radiation field) to dense, molecularclouds and eventually protostars. Thanks to the mechanical action of evolvedstars (winds, supernovae), molecular clouds are dispersed and restored to the dif-fuse ISM. Intimately mixed to the gas, dust and in particular the small grains (ofradius a ≤ 10 nm, which include PAHs) play an important role in the thermo-dynamics, dynamics and chemistry of the ISM. Dust is also the main source ofextinction and thus controls the radiative transfer in interstellar clouds.

The primary effect of the interstellar lifecycle on dust is to change the abun-dance of small grains with respect to larger ones (radii from 10 to a few 100 nm)during fragmentation (by radiation or shocks in the diffuse ISM) and growth (bycoagulation in dense regions) episodes. In the vicinity of young stars or in shocksPAHs are observed to be depleted (Geers et al. 2006, 2009; Peeters et al. 2005;

1 Institut d’Astrophysique Spatiale, UMR 8617, University Paris-Sud 11, Bat. 121,91405 Orsay Cedex, France



Lebouteiller et al. 2007; see Acke et al.; Rho et al., Sandstrom et al. andSiebenmorgen this volume). PAH coagulation is suggested by observations ofmolecular clouds (Boulanger et al. 1994; Flagey et al. 2009). Conversely, PAHsmay be photoevaporated off larger grains when fresh molecular material is exposedto stellar UV radiation thus producing a limb brightening effect (Boulanger et al.1990; Bernard et al. 1993; Rapacioli et al. 2005; Berne et al. 2007; Compiegneet al. 2008; Velusamy & Langer 2008; Rapacioli et al. this volume).

In the following, we first review interstellar processes where PAHs play animportant role and discuss the observational evidence. The next section is devotedto a key property of PAHs: their charge distribution. A brief summary is thengiven.

2 Physical processes involving PAHs

Interstellar PAHs are abundant (they bear about 20% of the total carbon abun-dance in diffuse clouds: Joblin et al. 1992; Zubko et al. 2004; Draine & Li 2007)and are the smallest grains (0.4 to 1 nm in radius) of dust grains. Because ofthese properties PAHs play a peculiar role in the physics and chemistry of theISM. While discussing the key interstellar processes which involve PAHs, we willsee that most of them require a detailed knowledge of the charge distribution ofPAHs.

2.1 Photoelectric heating of the gas

Observations at 21 cm show that the diffuse atomic (HI) gas is at temperaturesbetween 100 and 104 K and in rough pressure equilibrium. To explain thesetemperatures, early models of the gas thermal state proposed that the HI gas washeated by cosmic rays (Field et al. 1969) but it was later realized that the flux ofenergetic nuclei was much (∼10 times) lower. Watson (1972) then proposed thatthe heating of HI gas may be due to the photoelectric effect on interstellar grains.

In this process, an UV photon ejects an electron from the grain: this first stepis often called photoemission. The kinetic energy carried away by the electron(typically 1 to 2 eV) is then rapidly deposited in the gas by inelastic collisions (inless than 10 years for proton densities nH > 1 cm−3). To trigger photoemission,the energy of the incident photon must satisfy hν > IP where IP is the ionizationpotential or photoelectric threshold which depends on the charge Z (algebraicnumber of electron charges carried by the PAH) and size a of the grain. Thephotoelectric effect is characterized by its probability or yield and its energeticefficiency. The yield YPE is the ratio of the photoemission cross-section to theabsorption cross-section and depends on the energy of the photon1. Let Ee bethe mean energy of the electron ejected, the photoelectric efficiency is given byε = YPE Ee/hνa with hνa the mean energy of photons absorbed by the grain.

1We will quote here YPE for hν = 10 eV, roughly the energy of the maximum photoemissionrate in the Mathis et al. radiation field.

L. Verstraete: PAHs and ISM Physics 417

Interstellar grains may also capture electrons from the gas leading to an energygas loss proportional to kBT per capture (see Draine 1978, Eq. (10)). This processis called grain recombination cooling. Its contribution, particularly important inwarm gas, must be subtracted from the photoelectric heating rate to obtain thenet photoelectric rate ΓPE (erg cm−3 s−1).

Watson (1973) further pointed out that high yields (YPE > 0.1) were expectedfor small grains because the mean free path of the electron on its way to thegrain surface becomes comparable to the grain size. Early models of photoelectricheating assumed YPE ∼ 0.1 and were successful in explaining the cold HI gas buthad difficulties in supporting the warm (T > 5000 K) gas because of the stronggrain recombination cooling (Draine 1978). The photoelectric effect on PAHsdoes not suffer of this problem because of the higher energy of the ejected electronimplying a larger cut-off temperature for the heating rate2 (Verstraete et al. 1990).Because of their small size, PAHs have a high photoelectric yield (YPE ≥ 0.4)and, in the neutral gas, a small mean charge both properties implying a highphotoelectric efficiency (e.g., Bakes & Tielens 1994 hereafter BT94)3. Currentmodels of photoelectric heating show that ε is maximum (a few percents) in thelimit of small size (a < 1 nm) and charge (Z < 1) (BT94; Weingartner & Draine2001 hereafter WD01).

Theoretical studies of the gas thermal balance have shown that the photoelec-tric effect is the dominant heating process over a wide range of conditions in gasdensity 0.1 ≤ nH ≤ 104 cm−3 and intensity of radiation field 1 ≤ G0 ≤ 106. Onthe other hand, the gas cools by collisional excitation of optically thin emissionlines of major species such as C+, O, H, H2 and CO, depending on the gas densityand temperature. The photoelectric heating thus controls the gas temperatureand hence the line emissivity and chemistry of interstellar clouds. Moreover, thephotoelectric heating also determines the conditions under which the gas becomesthermally unstable, a key process in understanding the structure and evolution ofthe ISM (Audit & Hennebelle 2005, 2010).

The hypothesis of gas photoelectric heating can be tested observationally. Atdensities nH > 1 cm−3, the gas cools rapidly (in less than 0.1 Myr) and can beassumed to be in thermal equilibrium, i.e., ΓPE = Λ where Λ is the gas coolingrate in erg cm−3 s−1. Observations of the cooling lines thus directly trace ΓPE .The latter can then be compared to the dust IR emission to determine whichgrain sizes contribute most to the heating and with which efficiency, in order tocheck model predictions. Over the last decade, such studies have been performedin different contexts. Habart et al. (2001) made a detailed study of an isolatedcloud with well-known properties (exciting radiation field, density profile). UsingISO-LWS observations of the major cooling lines ([CII]158 μm and [OI]63 μm)and extracting the emission of each dust population from IRAS data, they foundthat the gas cooling is well correlated with the IR emission of small grains and

2The temperature at which the photoelectric heating and recombination cooling rates cancel.3Electrostatic forces bring IP close to 13.6 eV when Z > 0.


that the photoelectric heating efficiency for PAHs was 3% and decreases for largergrains. Modelling the [CII]158 μm emission measured by ISO-LWS towards 9 high-galactic latitude translucent clouds (AV ≤ 5), Ingalls et al. (2002) found an overallphotoelectric heating efficiency of 4.3%. They also showed that the decrease of CIIemission at high dust emission is due to radiative transfer inside clouds (see alsoJuvela et al. 2003). More recently, Rubin et al. (2009) studied the 30 Doradusstar forming region in the Large Magellanic Cloud with Spitzer IRAC+MIPS (dustemission) and BICE data (CII emission). They also found that PAHs have thehighest photoelectric efficiency, about 5%. Finally, in external normal galaxies,a correlation has been found between the CII line and the 5-to-10 μm ISOCAMemission ascribed to PAHs (Helou et al. 2001). All these results are in goodagreement with model predictions of the photoelectric effect on dust grains andshow the dominant role of PAHs for the gas heating.

2.2 Screening of UV radiation

Aromatic species like PAHs show strong UV absorption in two broad bands cen-tered at 200 nm (π∗ ← π transitions) and 70 nm (σ∗ ← σ transitions). PAHsare thus good candidates to explain the 217 nm bump and the far-UV rise ofthe interstellar extinction curve as suggested by recent theoretical and laboratorystudies (see the review by Mulas et al. this volume and Cecchi-Pestellini et al.2008; Steglich et al. 2009). In particular, PAHs absorb efficiently dissociatingphotons (at λ ≤ 100 nm) thus controlling the molecular fraction and excitation ofinterstellar clouds. This UV screen will be enhanced if PAHs are more abundantat cloud boundaries as suggested by observations (the limb brightening effect, seeSect. 1). As discussed by Mulas et al. (this volume), the far-UV absorption ofPAHs depends on their charge state.

2.3 Ionization of dense gas

The delocalized π-electrons of PAHs are easily polarizable (Omont 1986). Thus dueto their high polarizability and abundance PAHs can efficiently capture electronsand, in shielded interstellar regions, PAH anions may be as abundant as electrons(x = ne/nH ∼ 10−7). Charge exchange between positive ions generated by cosmicrays in the gas and PAH anions quenches the ion-molecule chemistry and mayreconcile observations and models (Bakes & Tielens 1998; Flower et al. 2007;Wakelam & Herbst 2009; Goicoechea et al. 2009). The role of PAHs in interstellarchemistry is discussed by Bierbaum et al. (this volume). PAH anions may alsoinduce recombination of CII, hence leading to warmer gas in diffuse clouds (Wolfireet al. 2008).

Charged PAHs are also important for the ion-neutral coupling and the dynam-ics. In magnetic C-shocks, PAHs may produce a shortage of electrons leading tothe decoupling of larger grains in the shock tail. At a given density, this will in-crease the Alfven velocity in the shock and its emissivity by the ion-neutral drift(Ciolek et al. 2002, 2004; Guillet et al. 2007). In the context of protostars, the


role of PAHs in braking the collapse has been studied by Ciolek and collaborators.At cloud boundaries where UV photons are less shielded, PAHs may carry positiveas well as negative charges (see Sect. 3): this could accelerate the coagulation ofPAHs together as suggested by the observations of Flagey et al. (2009).

2.4 H2 formation

It has long been known that, at ISM densities, H2 cannot form efficiently in the gasphase. In 1963, Gould & Salpeter proposed that H2 may form by catalytic reactionon interstellar grains where an H atom bound to the grain surface recombineswith another H. Molecular hydrogen has a profound impact on the physics andchemistry of the ISM: it efficiently cools the gas (Le Bourlot et al. 1999) and actsas a shield to photodissociating UV photons, initiating the formation of CO andof chemistry. In spite of many theoretical and experimental studies, the detailedmechanism by which H2 forms is complex and still debated because it depends onthe structure and composition of the grains (surface roughness, presence of chemi-or physisorption sites for H) and involves the issue of distributing the 4.5 eVbonding energy of the molecule between excitation energy (grain, H2 vibrationalor rotational) and H2 kinetic energy (see Bierbaum et al., Thrower et al. thisvolume). The rate coefficient for H2 formation is observed to vary from a fewtenths to a few times the standard value of Jura (1975), R0 = 3 × 10−17 cm3 s−1

(Gry et al. 2002; Browning et al., Habart et al. 2003; Gillmon et al. 2006; Thiet al. 2009). In addition, the observations of H2 emission in excited regions favorformation routes involving chemisorbed sites for H (Habart et al. 2004).

PAHs are attractive candidates for the formation of H2 because they repre-sent two thirds of the total grain surface, e.g., Habart et al. 2004) and carrychemically bonded H atoms. Observations indicate an expected trend between H2

and PAH integrated emissions (Habart et al. 2004) but the corresponding emis-sivity profiles are not always correlated, (Habart et al. 2003; Berne et al. 2009;Velusamy & Langer 2008). These results may be related to the radiative transferof UV photons across cloud borders with different density profiles. Shock excitedH2 emission does not suffer of this limitation and should represent a better test(Guillard et al. 2010). Analysing UV absorption data in diffuse clouds, Wolfireet al. (2008) modelled in a consistent way the C/C+ and HI/H2 transitions withPAH assisted recombination of C+ and photoelectric heating. Their results alsoindicate a correlation between the PAH abundance and the H2 formation rate.

2.5 Spinning dust emission and microwave polarization

At centimetric wavelengths (around 30 GHz), an emission component discoveredby Kogut et al. (1996) in COBE data was later shown to correlate with the emis-sion of large dust grains (Leitch et al.; de Oliveira-Costa et al. 1997). Unexpectedin this spectral range where the free-free, synchrotron and large grain emissionsusually dominate, this component was dubbed anomalous emission. Due to thismixing, the existence of anomalous emission has long been debated and it is only


recently that it has been extracted in all-sky WMAP data thanks to the constraintson polarization (Miville-Deschenes et al. 2008). Understanding this emission isparticularly important because it appears in a spectral region where fluctuationsof the cosmic microwave background are best measured. It is now further tracedby the measurements of the Planck mission.

Several emission mechanisms have been proposed to explain the anomalousemission. Draine & Lazarian (1998a) considered the rotational emission of PAHs(or spinning dust emission). Draine & Lazarian (1999) further proposed magneticdipole emission of grains with magnetic inclusions and Jones (2009) proposedlow-energy transitions arising in disordered grains. The latter two mechanismsinvolve large grains (a ∼ 0.1 μm). Recent analysis of observations indicate thatthe anomalous emission correlates better with the mid-IR emission due to PAHs(Casassus et al. 2006, 2009; Ysard et al. 2010) than with the emission of largergrains. Another hint comes from polarization: the anomalous emission appears tobe weakly polarized (Battistelli et al. 2006) as expected for small grains (Lazarian& Draine 2000; Martin et al. 2007) thus favoring the spinning dust explanation4.These results have stimulated new theoretical studies of PAH rotational dynamicsin the ISM (Ali-Haımoud et al. 2009; Ysard & Verstraete; Hoang et al.; Silsbeeet al. 2010) which confirmed the main features of the Draine & Lazarian (1998b)model. Namely, the PAH rotation is mostly controlled by gas-grain collisions and,in irradiated regions, also by their vibrational (IR) emission. In addition, theseprocesses depend on the PAH charge.

The anomalous emission may thus trace the PAH abundance and provide newconstraints on PAHs such as their permanent electric dipole moment. Further on,the contribution of spinning PAHs could reduce the polarization fraction measuredin the microwave by Planck. Quantitative modelling of this emission componentis necessary for the analysis of the Planck polarization data in the cosmologicalcontext but also for ISM studies (magnetic field).

3 The charge distribution of PAHs

Most of the processes discussed in the previous section require a detailed knowledgeof the PAH charge distribution. This is also true for the spectroscopic propertiesof PAHs, from the IR to the UV (Pauzat, Pino et al., Mulas et al., Oomens thisvolume). We therefore review the current state-of-the-art in modelling the chargeof interstellar PAHs.

In the ISM PAHs become charged because of photoemission or collisions withcharged species. Excluding dark clouds (where PAHs may have coagulated onlarger grains), the latter case involves the most abundant species e−, H+ andC+. Since the characteristic times for charge fluctuations are short, a stationarydescription is often adopted (see Ivlev et al. 2010 however for a discussion of the

4Conversely, large grains are known to polarize light in emission (FIR to submm) and ex-tinction (IR to UV), indicating that grains are not spherical and aligned with the magneticfield.


Fig. 1. a) Rate coefficient at 300 K as a function of size (in number of C atoms per

molecule) for the electronic recombination of PAH+ (adapted from Tielens 2008). The

data points are from Biennier et al. (2006). The bottom solid line shows the law adopted

by WD01 and the top solid line that of BT94 (se = 1). The dashed line shows the

se = 0.1-case. b) Sticking coefficient for electronic capture on a neutral PAH as a

function of size (adapted from WD01, note that triangles at large sizes are for C60 and

C70). Solid lines are as in a). The dashed line shows the law of Dartois & d’Hendecourt

(1997).

effect of charge fluctuations). The charge distribution f(Z) is then derived fromthe balance equation (BT94):

f(Z) [JPE(Z) + Ji(Z)] = f(Z + 1)Je(Z + 1) (3.1)

written for a given PAH size and where the J-terms represent charge currents.Estimates of the latter require the knowledge of rate coefficients which dependson intrinsic quantities such as the photoelectric yield, the threshold IP and theelectron capture rate coefficient (or sticking coefficient) (see Sect. 2.1). Two typesof models have been developed: (i) molecular models using the properties measuredin the laboratory on small, real molecules (Allain et al., Salama et al. 1996; Le Pageet al. 2001, 2003) and, (ii) astrophysical models treating PAHs as generic specieswhose properties follow size trends based on laboratory measurements (BT94;Dartois & d’Hendecourt 1997; WD01; van Hoof et al. 2004). Being more widelyapplied, we focus here on the models of BT94 and WD01.

The photoelectric (or photoionization) yield for PAHs has been measured on afew, small (containing less than 25C atoms) PAHs (Verstraete et al. 1990; Jochimset al. 1996). The expressions used by BT94 and WD01 provide a good descriptionof these data and extrapolate well to measurements on larger (a ≥ 100 nm) grains(Abbas et al. 2006). The functions for the photoelectric threshold IP (Z, a) used inboth models are also similar with the exception of the electron affinity (Z = −1),


Fig. 2. Temperature profile as a function of optical depth for two PDRs of a) low and

b) high excitation. In both panels, the bottom (top) solid line shows the WD01 (BT94)

case respectively.

somewhat overestimated by BT94. It is worth noting that these thresholds do notapply to fullerenes such as C60, C70 (BT94).

The situation is different for the sticking coefficient (or probability) se of theelectronic capture: WD01 use a size dependent coefficient whereas BT94 assumese = 1 for all sizes. These two prescriptions are compared to laboratory data inFigure 1 as a function of molecular size. Note that for interstellar PAHs, one hasNC > 15 (Le Page et al. 2003). The prescription of BT94 leads to a more efficientelectronic capture and hence more abundant neutral or anionic species. We notethat the laboratory measurements for PAH+ recombination fall between the BT94and WD01 prescriptions which can therefore be seen as extreme cases. Anotherconsequence is that the photoelectric heating rate of BT94 is stronger than that ofWD01 because PAHs are less positively charged in the former case (see Sect. 2.1).We illustrate this in Figure 2 by comparing the gas temperature across two pho-todissociated interfaces or photon-dominated regions (PDR) computed with theMeudon PDR code (Le Petit et al. 2006). The physics and chemistry of suchregions depend on the gas density nH and on the intensity of the UV radiationfield represented by G0

5. We ran isochoric models irradiated on a single side(see Fig. 2) and with the following gas abundances (in ppm): [C]=130, [O]=320,[N]=75, [Mg]=35, [Fe]=28, [Si]=1. As expected, in the BT94 case the mean charge

5G0 is a scaling factor for the energy density of the radiation field integrated between 6 and13.6 eV; G0 = 1 corresponds to the standard radiation field of Mathis et al. (1983).


Fig. 3. Line emissivity profiles as a function of optical depth in two PDRs. As in Fig. 2,

the bottom (top) solid and dashed lines represent the case of WD01 (BT94). The solid

line shows the [OI]63 μm emissivity and the dashed line shows the emissivity in the first

4 rotational lines of H2. For clarity the H2 emissivities have been multiplied by 50.

< Z > of a 20C PAH is shifted by −0.5 to −1 with respect to the WD01 case6.The photoelectric heating rate of BT94 is correspondingly higher (2 to 4 timeshigher than WD01 for the present G0 = 10 and 103 PDRs respectively). Evenfor conditions of low excitation (G0 = 10), the BT94 model produces significantcolumn densities of warm gas. In an edge-on geometry, the line emissivity (in ergcm−3 s−1) of major gas coolants is significantly affected as illustrated in Figure 3.Integrated line fluxes (face-on geometry) are less affected (below 30% variations).Such differences will lead to quite different diagnostics of the physical conditions(intensity of the UV radiation field, gas density and temperature) and emphasizethe need for more laboratory measurements of the electronic capture on astrophys-ically relevant PAHs (containing more than 20C). Such measurements should alsobe carried at different temperatures to check that the classical limit often usedprovides a reasonable description7.

Thanks to the contrasted behaviour of the CC-to-CH ratio of the vibrationalband intensities in neutral and ionized PAHs (e.g., Bauschlicher et al. 2008 andPauzat this volume), the charge distribution of PAHs can be constrained frommid-IR observations. Flagey et al. (2006) found that, in the diffuse ISM, 40%of PAHs are ionized. Analyzing a sample of galactic and extragalactic excited

6With the WD01 model, the mean charge is +0.5 (+0.9) for the G0 = 10 (103) PDR respec-tively.

7Some data suggest a behaviour departing from the classical T 1/2 (Biennier et al. 2006).


regions, Galliano et al. (2008 and Galliano this volume) found the expected trendbetween the fraction of ionized PAHs and excitation conditions.

4 Summary

PAHs have a strong impact on the ISM: they dominate the gas heating in irradiatedregions and carry a significant fraction of the dust surface, a potential site forthe formation of H2 and other molecules. In dense clouds, PAHs may becomean important sink for electrons and control the ion-molecule chemistry as wellas the dynamics of neutrals relative to ionized species in presence of a magneticfield. PAHs are therefore key to understanding the evolution of interstellar gas.Most of the processes involving PAHs depend on their charge state and this isalso true of their spectroscopical properties. Quantitative state-of-the-art modelsof the PAH charge are limited by the lack of data on the rate coefficient for theelectronic capture on large species containing more than 20 carbon atoms (see alsoMontillaud et al. this volume). This leads to significant uncertainties in derivingthe physical conditions and line emission of interstellar clouds. More laboratorymeasurements are needed in this area.

The author is grateful to Vincent Guillet and Jacques Le Bourlot for useful discussions and helpwhile preparing this review.

References

Abbas, M.M., et al., 2006, ApJ, 645, 324

Ali-Haımoud, Y., Hirata, C.M., & Dickinson, C., 2009, MNRAS, 395, 1055


Audit, E., & Hennebelle, P., 2005, A&A, 433, 1

Audit, E., & Hennebelle, P., 2010, A&A, 511, A76

Bakes, E.L.O., & Tielens, A.G.G.M., 1994, ApJ, 427, 822 (BT94)


Bauschlicher, C.W., Jr., Peeters, E., & Allamandola, L.J., 2008, ApJ, 678, 316

Battistelli, E.S., Rebolo, R., Rubino-Martın, J.A., et al., 2006, ApJ, 645, L141

Bernard, J.P., Boulanger, F., & Puget, J.L., 1993, A&A, 277, 609

Berne, O., et al., 2007, A&A, 469, 575

Berne, O., Fuente, A., Goicoechea, J.R., et al., 2009, ApJ, 706, L160

Biennier, L., Alsayed-Ali, M., Foutel-Richard, A., et al., 2006, Chem. Evolu. Universe,Faraday Discuss., 133, 289

Boulanger, F., Falgarone, E., Puget, J.L., & Helou, G., 1990, ApJ, 364, 136

Boulanger, F., Prevot, M.L., & Gry, C., 1994, A&A, 284, 956

Browning, M.K., Tumlinson, J., & Shull, J.M., 2003, ApJ, 582, 810

Casassus, S., Cabrera, G.F., Forster, F., et al., 2006, ApJ, 639, 951

Casassus, S., et al., 2008, MNRAS, 391, 1075


Ciolek, G.E., & Roberge, W.G., 2002, ApJ, 567, 947

Ciolek, G.E., Roberge, W.G., & Mouschovias, T.C., 2004, ApJ, 610, 781

Compiegne, M., Abergel, A., Verstraete, L., & Habart, E., 2008, A&A, 491, 797

Dartois, E., & D’Hendecourt, L., 1997, A&A, 323, 534

de Oliveira-Costa, A., Kogut, A., Devlin, M.J., et al., 1997, ApJ, 482, L17

Draine, B.T., 1978, ApJS, 36, 595

Draine, B.T., & Lazarian, A., 1998a, ApJ, 494, L19

Draine, B.T., & Lazarian, A., 1998b, ApJ, 508, 157

Draine, B.T., & Li, A., 2007, ApJ, 657, 810

Field, G.B., Goldsmith, D.W., & Habing, H.J., 1969, ApJ, 155, L149

Flagey, N., Boulanger, F., Verstraete, L., et al., 2006, A&A, 453, 969

Flagey, N., et al., 2009, ApJ, 701, 1450

Flower, D.R., Pineau Des Forets, G., & Walmsley, C.M., 2007, A&A, 474, 923

Geers, V.C., et al., 2006, A&A, 459, 545

Geers, V.C., van Dishoeck, E.F., Pontoppidan, K.M., et al., 2009, A&A, 495, 837

Gillmon, K., Shull, J.M., Tumlinson, J., & Danforth, C., 2006, ApJ, 636, 891

Goicoechea, J.R., Pety, J., Gerin, M., Hily-Blant, P., & Le Bourlot, J., 2009, A&A, 498,771

Gould, R.J., & Salpeter, E.E., 1963, ApJ, 138, 393

Gry, C., Boulanger, F., Nehme, C., et al., 2002, A&A, 391, 675

Guillard, P., Boulanger, F., Cluver, M.E., et al., 2010 [arXiv:1004.0677]

Guillet, V., Pineau Des Forets, G., & Jones, A.P., 2007, A&A, 476, 263


Habart, E., Boulanger, F., Verstraete, L., et al., 2003, A&A, 397, 623

Habart, E., Boulanger, F., Verstraete, L., Walmsley, C.M., & Pineau des Forets, G.,2004, A&A, 414, 531

Hoang, T., Draine, B.T., & Lazarian, A., 2010, ApJ, 715, 1462

Ingalls, J.G., Reach, W.T., & Bania, T.M., 2002, ApJ, 579, 289

Ivlev, A.V., Lazarian, A., Tsytovich, V.N., et al., 2010 [arXiv:1003.2633]

Jochims, H.W., Baumgaertel, H., & Leach, S., 1996, A&A, 314, 1003

Jones, A.P., 2009, A&A, 506, 797

Jura, M., 1975, ApJ, 197, 575

Juvela, M., Padoan, P., & Jimenez, R., 2003, ApJ, 591, 258

Kogut, A., Banday, A.J., Bennett, C.L., et al., 1996, ApJ, 460, 1

Lazarian, A., & Draine, B.T., 2000, ApJ, 536, L15

Le Bourlot, J., Pineau des Forets, G., & Flower, D.R., 1999, MNRAS, 305, 802

Lebouteiller, V., Brandl, B., Bernard-Salas, J., Devost, D., & Houck, J.R., 2007, ApJ,665, 390

Leitch, E.M., Readhead, A.C.S., Pearson, T.J., & Myers, S.T., 1997, ApJ, 486, L23





Martin, P.G., 2007, EAS Publications Series, 23, 165

Mathis, J.S., Mezger, P.G., & Panagia, N., 1983, A&A, 128, 212

Miville-Deschenes, M.-A., Ysard, N., Lavabre, A., et al., 2008, A&A, 490, 1093

Omont, A., 1986, A&A, 164, 159

Peeters, E., Martın-Hernandez, N.L., Rodrıguez-Fernandez, N.J., & Tielens, X., 2005,Sp. Science Rev., 119, 273


Rubin, D., et al., 2009, A&A, 494, 647


Silsbee, K., Ali-Haimoud, Y., & Hirata, C.M., 2010 [arXiv:1003.4732]


Thi, W.-F., van Dishoeck, E.F., Bell, T., Viti, S., & Black, J., 2009, MNRAS, 400, 622

van Hoof, P.A.M., Weingartner, J.C., Martin, P.G., Volk, K., & Ferland, G.J., 2004,MNRAS, 350, 1330

Velusamy, T., & Langer, W.D., 2008, AJ., 136, 602

Verstraete, L., Leger, A., d’Hendecourt, L., Defourneau, D., & Dutuit, O., 1990, A&A,237, 436


Watson, W.D., 1972, ApJ, 176, 103

Watson, W.D., 1973, J. Opt. Soc. Amer. (1917-1983), 63, 164

Weingartner, J.C., & Draine, B.T., 2001, ApJS, 134, 263 (WD01)

Wolfire, M.G., Tielens, A.G.G.M., Hollenbach, D., & Kaufman, M.J., 2008, ApJ, 680,384

Ysard, N., Miville-Deschenes, M.A., & Verstraete, L., 2010, A&A, 509, L1




PAHS AND THE CHEMISTRY OF THE ISM

V.M. Bierbaum1, V. Le Page1 and T.P. Snow2

Abstract. This review describes five categories of the reactions of poly-cyclic aromatic hydrocarbons, including photochemistry, electron at-tachment/detachment, recombination processes, radical reactions, andion-neutral chemistry. For each class of reaction, an overview of thestudies and their general results are presented, as well as references tothe literature. The thermochemistry of PAHs and relevant species is de-scribed, including bond dissociation energies, ionization energies, elec-tron affinities, basicities, acidities, and the interrelationships of thesequantities. Modeling of the chemistry of PAHs and their ions is dis-cussed for both diffuse and dark clouds. The role of PAH cations in thecatalytic formation of molecular hydrogen is considered. Finally, thisreview concludes with a discussion of current challenges in the chemicalcharacterization of PAHs, and a perspective for future studies.

1 Introduction

The history of polycyclic aromatic hydrocarbons (PAHs) in the universe has been afascinating saga, beginning with the ground-breaking proposal of their importancein the cosmos twenty-five years ago (Leger & Puget 1984; Allamandola et al. 1985).In the intervening years, many scientists have contributed to the unfolding of thisstory through astronomical observations, laboratory spectroscopy and chemistry,computational characterization, and detailed modeling of astrophysical environ-ments. The integration of these methods and their results has now confirmed thepresence of PAHs in many locations throughout the universe. This review willfocus on the chemistry of PAHs in the interstellar medium (ISM)(Tielens 2008;Snow & Bierbaum 2008). Chemistry can be defined as the interactions of elementsand compounds with photons, ions, and neutral species, resulting in the formationand dissociation of bonds, the gain and loss of electrons, and the synthesis of therich chemical structure of the universe. Thus, although chemistry represents only

1 Department of Chemistry & Biochemistry; University of Colorado, Boulder, CO 80309,USA2 Department of Astrophysical & Planetary Science; University of Colorado, Boulder,CO 80309, USA



one facet of the complex subject of PAHs in the universe, it is nevertheless anextremely broad and diverse topic, and only a small portion of the literature canbe highlighted in this review. Gas phase processes will be discussed at the level ofdiscrete molecules; the chemistry of clusters and grains forms the topics of othermanuscripts in this volume.

2 Reactions of PAHs

There are several valuable compilations of gas phase reactions: the UMIST Data-base for Astrochemistry (http://www.udfa.net/); the Ohio State Network(http://www.physics.ohio-state.edu/~eric/research_files/osu_01_2009);KIDA, the Kinetic Database for Astrochemistry (http://kida.obs.u-bordeaux1.fr/); and the NIST Chemical Kinetics Database (http://kinetics.nist.gov/kinetics/index.jsp).

2.1 Photochemistry

Photochemistry involves chemical transformations resulting from the interactionof light with matter. Turro et al. (2010) have provided a comprehensive review ofthe molecular photochemistry of organic molecules that involves a wide array ofreactions. This paper will focus on gas phase processes likely to occur in the ISM:the photoionization and dissociation of molecules to form reactive species.

Using naphthalene as an example, photon absorption can result in ionizationto form the parent molecular cation; alternatively dissociation can accompanyionization with the loss of hydrogen atom, a hydrogen molecule, or extrusion of asmall hydrocarbon such as acetylene.

C10H8 + hν −→ C10H+8 + e−

−→ C10H+7 + H + e−

−→ C10H+6 + H2 + e−

−→ C8H+6 + C2H2 + e−

There have been many experimental and theoretical studies of the photochemistryof PAHs and their cations (Verstraete et al. 1990; Jochims et al. 1996; Gotkiset al. 1993; Banisaukas et al. 2004); for example, a comprehensive explorationof the photostability of 33 compounds was carried out using monochromatizedsynchrotron radiation (Jochims et al. 1999). Results from these various studiesindicate that there is a rapidly increasing ionization yield above the ionizationenergy of the PAH, and rules for determining the quantum yield as a function ofwavelength have been proposed. These studies have determined the fragmentationchannels and the internal energy required for fragmentation; in general, the lossrates for H from alkyl groups exceed those of H from aromatics, which exceedcarbon loss as in the form of acetylene.

Photodissociation is being increasingly implemented in Fourier transform ioncyclotron resonance (FT-ICR) experiments to characterize the visible and infraredspectroscopy of PAH cations. The electronic spectra of the pyrene cation, the

V.M. Bierbaum et al.: PAH Chemistry in the ISM 429

1-methylpyrene cation, the coronene cation and its dehydrogenated derivativehave been measured in the PIRENEA experiment using a mid-band OPO laser(Useli-Bacchitta et al. 2010). The infrared spectra of protonated naphthalene,azulene, anthracene, tetracene, pentacene, and coronene have been determinedusing a free electron laser (Zhao et al. 2009; Knorke et al. 2009).

2.2 Electron attachment/detachment

Negative ion formation resulting from low energy (0–15 eV) electron impact ofPAHs has been studied in molecular beam experiments (Tobita et al. 1992). Theelectron detachment cross section is a strong function of the electron affinity of thePAH, which generally increases with size. Thus, for naphthalene, phenanthrene,and triphenylene (which have a negative or only slightly positive electron affinity)no long-lived parent anions (M−) were observed. The transient negative ion au-todetaches before stabilization by photon emission (Tielens 2005). However, formolecules where EA > 0.5 eV, parent negative ions were detected; these com-pounds include azulene, anthracene, pyrene, acridine, fluoranthene, perylene, andtetracene. For all ten compounds studied, dehydrogenated anions (M-H)− wereobserved at onset electron energies of about 6 eV (Tobita et al. 1992).

The room temperature rate coefficient for electron attachment to anthracenewas measured in a flowing afterglow-Langmuir probe (FALP) experiment to be 1 ×10−9 cm3 s−1 (Canosa et al. 1994). This relatively low value, corresponding to anelectron sticking coefficient of only 10−4, suggests the existence of an energy barrierfor this process or the absence of accessible negative ion states for attachment.However, the electron attachment rate coefficient is predicted to have a valueabout 10−6 cm3 s−1 for PAHs with 30 carbon atoms (Omont 1986; Allamandolaet al. 1989). FALP studies indicate that both C60 and C70 efficiently attachelectrons above an electron energy of 0.2 eV (Smith & Spanel 1996).

Photodetachment of PAH anions, with energy resolution of the photoelectrons,has been studied for anthracene and perylene (Schiedt & Weinkauf 1997a, 1997b),and for chrysene (Tschurl & Boesl 2006). The electron affinities, vibrational fre-quencies, and singlet-triplet energy gaps for the neutral molecules can be deter-mined. These relatively limited studies indicate that additional work is needed tofully understand the electron attachment/detachment processes of PAHs.

2.3 Recombination

The recombination of molecular cations with electrons or with negative ions re-sults in neutralization and dissociation. A variety of experimental techniques,including merged beams, ion storage rings, afterglow techniques, and shock-tubemethods, have contributed a wealth of data on the recombination of electrons witha variety of cations, from simple diatomics to polyatomic species. Larsson & Orel(2008) have provided an excellent overview of these processes. Many reaction rateconstants have been determined, and the temperature dependence and reactionproducts have been studied for several cations.


However, studies of ion-electron recombination of PAH cations are relativelysparse. Researchers at Rennes University (Biennier et al. 2006) have recentlyused the flowing afterglow/photoions method and the FALP technique to mea-sure the room temperature recombination rate coefficients of seven PAH cations:naphthalene [0.3 (± 0.1) × 10−6 cm3 s−1], azulene [1.1 (± 0.3) × 10−6 cm3 s−1],acenaphthene [0.5 (± 0.2) × 10−6 cm3 s−1], anthracene [2.4 (± 0.8) × 10−6 cm3

s−1], phenanthrene [1.7 (± 0.5) × 10−6 cm3 s−1], fluoranthene [3.0 (± 0.9) ×10−6 cm3 s−1], and pyrene [4.1 (± 1.2) × 10−6 cm3 s−1]. The results indicatean increased reactivity as the number of carbon atoms increases; although thereactions are rapid, they occur below the geometrical limit. Product distributionswere not determined; however, these reactions can occur dissociatively to extrude ahydrogen atom or small molecule, or non-dissociatively, especially for larger PAHs.

Hamberg et al. (see elsewhere in this volume) report the temperature depen-dence of the recombination of the benzene cation and of the protonated benzenecation, using the CRYRING. For C6D6

+, k(T) = 1.25 × 10−6 (T/300)−0.69 cm3 s−1,and for C6D+

7 , k(T) = 2.0 × 10−6 (T/300)−0.83 cm3 s−1. These researchers findthat the aromatic ring remains largely intact.

Future studies of the dissociative recombination of larger PAH cations, theirtemperature dependence, and their reaction products are highly desirable. In addi-tion, while positive ion-negative ion reactions have been studied for simple systems,no measurements exist for PAH ions. Due to the positive electron affinity of mostPAH molecules, these processes will be less exothermic than the correspondingion-electron recombination, and therefore less dissociation may occur. The devel-opment of DESIREE (Schmidt et al. 2008), with two cryogenic ion-storage rings,will make these novel and important studies possible.

2.4 Radical reactions

Radical-neutral reactions may be important in the interstellar medium even atlow temperatures (Ekern et al. 2007). Recently, Rowe and coworkers have uti-lized the CRESU apparatus to measure the rate constants for radicals reactingwith anthracene between 58-470 K. For OH radical (Goulay et al. 2005), thereaction is rapid and increases as the temperature is lowered: k = 1.12 × 10−10

(T/300)−0.46 cm3 s−1; formation of the adduct is inferred. For CH radical (Goulayet al. 2006a), the reaction exhibits a slight positive temperature dependence: k =(3.32 ± 1.00) × 10−10 (T/298)(0.46±0.14) cm3 s−1; however, even at the lowesttemperature, the reactivity remains high. The products of the reaction were notdetermined.

There is an extensive literature describing the synthesis of PAHs under var-ious conditions (Tielens 2008); in some environments the reactions of benzenewith carbon-containing radicals provide a probable mechanism for the formationof PAHs (see Cherchneff, this volume). For example, the role of the CH3 radi-cal has been explored, and found to be important in PAH growth (Shukla et al.2010). The rate coefficient for the reaction of the C2H radical with benzene has


recently been characterized between 105 and 298 K; this value [k = (3.28 ± 1.0) ×10−10 (T/298)−(0.18±0.18) cm3 s−1] indicates that the reaction has no barrier andmay play a central role in the formation of large molecules (Goulay & Leone2006b). Additional T-dependent studies of the reactions of radicals with aromaticmolecules are clearly warranted.

An especially challenging aspect of the chemistry of PAHs is the existence ofisomeric forms and the resulting differences in reactivity and thermodynamics.There is a recent innovative study of Zwier & co-workers (2010) that explores site-specific thermochemistry of PAH radicals. These workers have determined bondenergies and ionization potentials for the C10H9 isomeric radicals, 1-hydronaphthyland 2-hydronaphthyl, from their jet cooled vibronic spectra. These studies revealthe intricacies and the richness of the chemistry of the simplest PAH.

IP (1HN) = 6.570 eV

BDE = 121.2 kJ/mol

IP (2HN) = 6.487 eV

BDE = 103.6 kJ/mol

2.5 Ion-neutral reactions

The area of gas phase ion chemistry represents an extremely diverse and complexsubject of study. Reactions of cations and anions, both simple and complex, canoccur with atoms, radicals, and molecules; a multitude of processes are observed,including charge transfer, proton transfer, addition, carbon insertion/loss, hydro-gen addition/abstraction, isotope exchange, and associative detachment. Quanti-tative rate constants, their temperature dependence, and the product distributionsare needed for models of interstellar chemistry (Herbst 2001; Smith 2006). Thisreview will briefly consider three categories of reactions, those involving smallorganic species, fullerenes, and PAHs.

The reactions of small organic ions, CwHxNyOz±, with a host of neutral reac-

tants have been characterized by many research groups, using a variety of tech-niques including flow tubes, ICR instruments, and ion traps. There are severalexcellent reviews of the field (Smith & Adams 1988; Bohme 2000; Uggerud 2003;Gronert 2005; Nibbering 2006; Gerlich & Smith 2006). These sequential reactionsprovide the synthesis of complex species, including PAHs. Although past work hasfocused on positive ions, anions have recently been detected in molecular clouds;thus, the research in our laboratory is currently centered on negative ion chemistryof small carbon-containing species (Eichelberger et al. 2007; Yang et al. 2010).

Bohme has utilized flow tube mass spectrometry to explore the chemistry of themono-, di-, and trications of buckminsterfullerene and several substituted cations(Bohme 2009). These ions show remarkable reactivity with a wide variety of


reagents, which form covalent bonds and thereby generate derivatized fullerenes.Although these processes involve three-body stabilization under the experimentalconditions, the radiative processes are likely to be rapid. In analogy, large PAHions are expected to serve as efficient substrates for addition of many atomic ormolecular species.

Unfortunately, experimental studies of the ion chemistry of PAHs are rela-tively limited. Our work (Snow et al. 1998; Le Page et al. 1999a; Le Pageet al. 1999b; Betts et al. 2006) has employed the flowing afterglow-selected ionflow tube (FA-SIFT) technique to explore the chemistry of the parent molecularcations of benzene, naphthalene, pyrene, and coronene (C6H6

+, C10H8+, C16H10

+,C24H12

+), the monodehydrogenated cations (C6H5+, C10H7

+, C16H9+), as well

as protonated benzene, naphthalene, and pyrene (C6H7+, C10H9

+, C16H11+). For

all of these ions, reactions with H2, H, N, and O were characterized; for some ions,reactions with other molecular reagents were also examined (CO, H2O, NH3, N2,O2, NO, CH3OH, CS2, (CH3)3N). Consider, for example, the reactions of C10H8

+:

C10H+8 + H2 −→ no reaction

C10H+8 + H −→ C10H+

9

C10H+8 + O −→ C10H8O+ (0.45)

−→ C9H+8 + CO (0.55)

C10H+8 + N −→ C10H8N+ (0.7)

−→ C9H+7 + HCN (0.3)

The parent molecular cation of naphthalene does not react with molecular hy-drogen. However, this radical cation reacts readily with the radical, H-atom, toform protonated naphthalene. In contrast, singlet cations (C6H+

5 , C10H+7 ) have

only low reactivity with H atom, but do react with H2 by addition with moderaterates. Intriguingly, C16H+

9 is computed to have a triplet ground state, and thision reacts efficiently with H atom but not with H2. Our studies suggest that mostPAH cations will hydrogenate with either atomic or molecular hydrogen.

Although the addition processes in flow tube experiments involve three-bodystabilization, it is expected that radiative stabilization will be facile in the ISM. Forexample, radiative association of molecular hydrogen with C6H+

5 and C10H+7 has

been observed (Le Page et al. 1999a; Ausloos et al. 1989) at low pressure in ICRinstruments. Since the radiative process C10H+

7 + H2 is efficient, the same is likelytrue for the C10H+

8 + H reaction; the computed energy content of the intermediatecomplex [C10H+

9 ] is almost identical for these processes, and therefore the complexlifetimes are very similar.

The reactions of C10H+8 with O-atoms and N-atoms illustrated above occur

both by association and extrusion of a small stable molecule. This latter channeldecreases as the size of the PAH cation increases; this pathway represents less than5% of the reaction for C16H+

10, and is not detectable for C24H+12. Thus, for large

PAH cations, only addition is observed.Table 1 summarizes the rate constants for reactions of four cations with atomic

reagents. For both H-atoms and O-atoms, the rate constants are large, and the


Table 1. Comparison of the Reactivity of PAH Parent Molecular Cations with Atomic

Reactants. Reaction rate constants are given in units of cm3 s−1.

PAH Cations H O N+

2.2 × 10−10 1.0 × 10−10 1.2 × 10−10

+1.9 × 10−10 1.0 × 10−10 2.3 × 10−11

+1.4 × 10−10 1.0 × 10−10 1.5 × 10−12

+

1.4 × 10−10 1.3 × 10−10 No reaction

reactivity remains high as the PAH increases in size; this suggests that large PAHcations, which might exist in the diffuse ISM, will also be reactive with theseatomic reagents. In contrast, the reactivity with N-atoms plummets, and becomesnegligible for large PAH cations.

3 Thermochemistry of PAHs

The chemical reactions of PAHs are intimately related to their thermochemistry.Photochemistry is linked with ionization energy, electron attachment/detachmentdepends on electron affinity, and radical reactivity reflects bond energies; ion-neutral reactions involve these parameters as well as gas phase acidity and basicityvalues. These quantities are interrelated through thermochemical cycles so thatvalues, which are not easily accessible experimentally, can be computed.

For example, the gas phase acidity of a molecule RH can be measured with flowtube techniques, and the electron affinity of R can be accurately determined withphotoelectron spectroscopy; these values can be combined with the well-knownionization energy of the hydrogen atom to yield the R-H bond dissociation energy.

RH −→ R− + H+ ΔacidH(RH)R− −→ R + e− EA(R)

H+ + e− −→ H − IE(H)

RH −→ R + H D(RH)

Figure 1 illustrates the relationship of this negative ion cycle with the protonaffinity cycle and the appearance energy cycle (Ervin 2001). A comprehensivecompilation of thermodynamic quantities for many PAHs is provided in the NISTChemistry WebBook (http://webbook.nist.gov/chemistry/) (NIST 2010).


Fig. 1. Thermochemical energy level diagram illustrating the relationship of the proton

affinity cycle, the negative ion cycle, and the appearance energy cycle (Ervin 2001).

4 Modeling of the chemistry of PAHs

4.1 Diffuse clouds

Le Page et al. (2001) have developed a model of the hydrogenation and chargestates of PAHs in diffuse clouds; these studies explore which species are likelyto survive in environments where the Diffuse Interstellar Bands are formed. Themajor processes include ionization and photodissociation in the interstellar UVfield, recombination of PAH cations, and chemistry of PAH cations with H2 andH atoms. The coupled kinetic equations are incorporated into a mathematicalframework. Experimental data are employed, and statistical theories are used toestimate properties that are not available from experiments. Other minor processesare considered, including electron attachment, photodetachment, photofragmen-tation with carbon loss, double ionization, and chemistry with minor species.

Twelve PAH systems were considered, ranging from benzene up to species con-taining 200 carbon atoms (Le Page et al. 2003). For a standard diffuse cloud,the total H density is set as 100 cm−3, the ratio of molecular hydrogen to atomichydrogen density is 0.5, the scaling of the UV field is unity, the temperature is100 K, and the ratio of electron density to total H density is 1.4 × 10−4; theeffects of varying these parameters were examined. Under most diffuse cloudconditions, the density of the cations exceeds that of neutrals, and the densityof anions is low. However, the hydrogenation state strongly depends on thesize of the molecule. Small PAHs with fewer than about 15–20 carbon atoms


are destroyed in most environments. Intermediate-size PAHs with 20–30 carbonatoms are stripped of peripheral hydrogen atoms, but retain their carbon skeleton.Larger PAHs primarily have normal hydrogen coverage, while very large PAHs maybe super-hydrogenated. These studies demonstrate that photodissociation andchemical reactions are critical processes governing the behavior of PAHs in thediffuse ISM.

Recently, Montillaud et al. (see elsewhere in this volume) have developedan extended model that includes PAH clusters and PAHs complexed with heavyatoms. Additionally, new reaction rate constants were employed. The first resultsfrom this model for coronene appear in this Proceedings volume.

4.2 Dense clouds

Wakelam & Herbst (2008) have recently added PAHs to their gas-phase network ofreactions to explore their role in the chemistry of cold dense cores. The processes ofradiative electron attachment to form PAH−, recombination between PAH anionsand small cations, and photodetachment of PAH anions by UV radiation were in-cluded. A pseudo-time-dependent model, which computes the chemical evolutionat a fixed temperature and density, was utilized for three different sets of elemen-tal abundances. The size and abundance of the PAHs were varied to determinethe impact of these values on the chemistry. The physical conditions includeda temperature of 10 K, a molecular hydrogen density of 104 cm−3, a cosmic-rayionization rate of 1.3 × 10−17 s−1, and a visual extinction of 10.

The abundance of PAH cations was found to be at least two orders of magni-tude smaller than the neutral or anionic PAH abundances. A dominant effect inthe calculations is that PAH anions replace electrons as the dominant carrier ofnegative charge; this factor reduces the overall ionization fraction because atomiccations recombine more rapidly with PAH− than with electrons. In addition, theabundance of many molecular ions and neutrals are strongly influenced by thepresence of PAHs. Inclusion of PAHs in the models significantly improves agree-ment with observational data for TMC-1, but not for L134N; this result suggeststhat there may be differences in the PAH properties for the two regions.

5 PAHs as catalysts in the synthesis of H2

Bauschlicher (1998) and Hirama and co-workers (2003) have suggested the possiblerole of PAH cations in the synthesis of molecular hydrogen:

PAH+ + H −→ PAH++H

PAH++H + H −→ PAH+ + H2

Net Reaction H + H −→ H2

PAH+ represents any normally hydrogenated PAH cation, while PAH++H repre-

sents a cation with an additional hydrogen atom. The first step involves theformation of protonated PAH by reaction of H-atom with the parent molecular


Fig. 2. Chemical mechanism for the catalytic formation of H2 in the interstellar medium

(Le Page et al. 2009).

cation; H-atom then abstracts the extra hydrogen to form H2. Since the cationsare formed and consumed, the net reaction is formation of a hydrogen moleculefrom two hydrogen atoms. While we have found the first process to be efficient,our experiments indicate that the reactions of protonated PAHs with H are slow.We have therefore proposed an alternate model.

Our proposed mechanism consists of four reactions:

PAH+ + H −→ PAH++H

PAH++H + e− −→ PAH−H + H2

PAH−H + H −→ PAH

PAH −→ PAH+ + e−

Net Reaction H + H −→ H2

The first step again involves the formation of protonated cations. These speciesthen undergo dissociative recombination with electrons to form the dehydrogenatedPAH (PAH−H) and molecular hydrogen; loss of H2 is feasible due to the low bind-ing energy of the two hydrogen atoms bonded to the same carbon. Our modelalso includes formation of PAH + H, which is not shown here for simplicity. Inthe third step, the dehydrogenated radical PAH−H combines with hydrogen atomto regenerate the neutral PAH; this molecule is then photoionized to complete thecycle. The complete scheme is summarized in Figure 2. The rate coefficient ofH2 production using this process compares well with a widely accepted value of3 × 10−17 cm3 s−1 for formation on grains in the diffuse ISM. It is inferred thatPAHs can contribute significantly to the formation of molecular hydrogen, espe-cially in environments where high grain temperatures impede the recombinationof hydrogen atoms on grain surfaces.


Recently, there have been additional novel proposals for the production of H2

in the ISM. Vala et al. (2009) have generated protonated 1,2-dihydronaphthaleneby electrospray ionization in an FT-ICR, coupled to an IR-tunable free electronlaser; infrared multiphoton dissociation studies, complemented by computations,demonstrate the formation of molecular hydrogen. Thrower et al. (see elsewhere inthis volume) have used both computational and experimental methods to exploresuperhydrogenated PAHs; using coronene as an example, these studies indicateefficient routes to H2 formation that may be important in the ISM.

6 Future perspectives

There have been major advances in the chemical characterization of PAHs duringthe last 25 years; nevertheless, there are critical questions remaining for each ofthe chemical processes discussed in this review. Powerful new developments inexperimental methods promise to offer new insights into PAH chemistry in thenear future; several of the most crucial areas will be briefly highlighted.

Recombination reactions. There are no experimental data on the neutralizationreactions of positive ions with negative ions for PAHs. The dual ion-beam stor-age ring instrument (DESIREE) (Schmidt et al. 2008) at Stockholm Universitywill enable the measurement of the cross-sections, rate coefficients, and productsdistributions for these critical reactions.

Reactions at low temperature and pressure. The CRESU technique (Smith2006) and variable temperature flow tube systems (Smith 1992) have providedexperimental data on a variety of ionic processes at low temperature. The recentdevelopment of an FT-ICR with cryogenic cooling of the reaction cell (PIRENEA)(Useli-Bacchitta et al. 2010) allows a wide range of ion chemistry experimentsunder conditions that approach the physical environment of interstellar space (P ≤10−10 mbar, Ttrap ∼ 35 K). In addition, two new experimental methods enablethe characterization of ion chemistry involving atomic reactants at low energies.These include the innovative atomic beam 22-pole ion trap (Gerlich & Smith 2006;Luca et al. 2006), which can approach temperatures as low as 10 K, and the novelmerged beam apparatus at the Columbia Astrophysics Laboratory (Bruhns et al.2010).

Characterization of reactants and products. Even the simplest polycyclic aro-matic hydrocarbon, naphthalene, possesses distinct chemical sites for protonation,deprotonation, dehydrogenation, addition, and other reactions. The resultingchemical complexity has largely been uncharacterized. New spectroscopic tools(Knorke et al. 2009; Sebree et al. 2010) offer a powerful approach for distinguish-ing isomeric forms of PAHs and their derivatives, including protonated, hydro-genated, and cyclic versus ring-opened species. Similarly, for many reactions, onlythe rate constants have been determined, and the identification of products andtheir branching fractions remains an important area for study.

Reactions of negative ions. In general, gas phase ion chemistry has focused onpositive ions, since these species are more readily generated, and PAH cations arethe probable form in the diffuse interstellar medium. However, the recognition that


PAH anions are important in dense clouds (Wakelam & Herbst 2008) underscoresthe need for exploring the chemistry of these species. Both electrospray ionization(Fenn et al. 1990) and chemical ionization (Vestal 2001) offer promising approachesfor the generation and study of PAH negative ions.

Large PAHs. The low volatility of PAH molecules has largely limited theirstudy in the gas phase to the smaller members of this family. However, modelingstudies indicate that moderate and large PAHs are likely to exist in the ISM. LaserInduced Acoustic Desorption (LIAD) (Shea et al. 2007) offers a new, versatileapproach for the study of PAHs and a host of other compounds. We are currentlycoupling a LIAD source to our FA-SIFT instrument to extend our previous workto large, astrophysically relevant PAHs.

In summary, the field of interstellar chemistry of PAHs has a promising andexciting future; it is fascinating to speculate about the marvelous results that willbe presented at the 50th anniversary symposium!

The authors are grateful to the National Aeronautics and Space Administration and the NationalScience Foundation for financial support. We appreciate the many contributions of our colleagues,students, and research associates. We thank Mr. Oscar Martinez Jr. and Mr. Nicholas Demaraisfor their help in preparation of this manuscript.

References



Ausloos, P., Lias, S.G., Buckley, T.J., & Rogers, E.E., 1989, Int. J. Mass Spectrom. IonProc., 92, 65

Banisaukas, J., Szczepanski, J., Eyler, J., & Vala, M., 2004, J. Phys. Chem. A, 108, 3723

Bauschlicher, C.W., 1998, ApJ, 509, L125

Betts, N.B., Stepanovic, M., Snow, T.P., & Bierbaum, V.M., 2006, ApJ, 651, L129

Biennier, L., Alsayed-Ali, M., Foutel-Richard, A., et al., 2006, Discuss. Faraday Soc.,133, 289

Bohme, D.K., 2000, Int. J. Mass Spectrom., 200, 97

Bohme, D.K., 2009, Mass Spectrom. Rev., 28, 672

Bruhns, H., Kreckel, H., Miller, K., et al., 2010, Rev. Sci. Instr., 81, 013112

Canosa, A., Parent, D.C., Pasquerault, D., et al., 1994, Chem. Phys. Lett., 228, 26

Eichelberger, B.R., Snow, T.P., Barckholtz, C., & Bierbaum, V.M., 2007, ApJ, 667, 1283

Ekern, S., Gomez, L., Szczepanski, J., & Vala, M., 2007, Chem. Phys., 331, 219

Ervin, K.M., 2001, Chem. Rev., 101, 391

Fenn, J.B., Mann, M., Meng, C.K., Wong, S.F., & Whitehouse, C.M., 1990, MassSpectrom. Rev., 9, 37

Gerlich, D., & Smith, M., 2006, Phys. Scr., 73, C25

Gotkis, Y., Oleinikova, M., Naor, M., & Lifshitz, C., 1993, J. Phys. Chem., 97, 12282

Goulay, F., Rebrion-Rowe, C., Garrec, J.L.L., et al., 2005, J. Chem. Phys., 122, 104308

Goulay, F., Rebrion-Rowe, C., Biennier, L., et al., 2006a, J. Phys. Chem. A, 110, 3132


Goulay, F., & Leone, S.R., 2006b, J. Phys. Chem. A, 110, 1875

Gronert, S., 2005, Mass Spectrom. Rev., 24, 100

Herbst, E., 2001, Chem. Soc. Rev., 30, 168

Hirama, M., Ishida, T., & Aihara, J.-I., 2003, J. Comput. Chem., 24, 1378

Jochims, H.W., Baumgartel, H., & Leach, S., 1996, A&A, 314, 1003

Jochims, H.W., Baumgartel, H., & Leach, S., 1999, ApJ, 512, 500

Knorke, H., Langer, J., Oomens, J., & Dopfer, O., 2009, ApJ, 706, L66

Larsson, M., & Orel, A.E., 2008, Dissociative Recombination of Molecular Ions(Cambridge University Press: Cambridge)

Le Page, V., Keheyan, Y., Snow, T.P., & Bierbaum, V.M., 1999a, J. Am. Chem. Soc.,121, 9435

Le Page, V., Keheyan, Y., Snow, T.P., & Bierbaum, V.M., 1999b, Int. J. Mass Spectrom.,185/186/187, 949





Linstrom, P.J., & Mallard, W.G. (eds.), NIST Chemistry WebBook, NIST StandardReference Database Number 69, National Institute of Standards and Technology,Gaithersburg, MD 20899, http://webbook.nist.gov

Luca, A., Borodi, G., & Gerlich, D., 2006, in “Interactions of Ions with Hydrogen Atoms.In Progress report in XXIV ICPEAC 2005”, ed. Colavecchia, F.D., Fainstein, P.D.,Fiol, J., Lima, M.A.P., Miraglia, J.E., Montenegro, E.C.. Rivarola, R.D. (Rosario,Argentina), 29

Nibbering, N.M.M., 2006, Mass Spectrom. Rev., 25, 962

Omont, A., 1986, A&A, 164, 159

Schiedt, J., & Weinkauf, R., 1997a, Chem. Phys. Lett., 266, 201

Schiedt, J., & Weinkauf, R., 1997b, Chem. Phys. Lett., 274, 18

Schmidt, H.T., Johansson, H.A.B., Thomas, R.D., et al., 2008, Int. J. Astrobiology, 7,205

Sebree, J.A., Kislov, V.V., Mebel, A.M., & Zwier, T.S., 2010, J. Phys. Chem. A, 114,6255

Shea, R.C., Habicht, S.C., Vaughn, W.E., & Kenttamaa, H.I., 2007, Anal. Chem., 79,2688

Shukla, B., Miyoshi, A., & Koshi, M., 2010, J. Am. Soc. Mass Spectrom., 21, 534

Smith, D., & Adams, N.G., 1988, Adv. At. Mol. Phys., 24, 1

Smith, D., 1992, Chem. Rev., 92, 1473

Smith, D., & Spanel, P., 1996, J. Phys. B: At. Mol. Opt. Phys., 29, 5199

Smith, I.W.M., 2006, Angew. Chem. Int. Ed., 45, 2842

Snow, T.P., Page, V.L., Keheyan, Y., & Bierbaum, V.M., 1998, Nature, 391, 259

Snow, T.P., & Bierbaum, V.M., 2008, Annu. Rev. Anal. Chem., 1, 229

Tielens, A.G.G.M., 2005, “The Physics and Chemistry of the Interstellar Medium”(Cambridge Univ. Press: Cambridge, UK)



Tobita, S., Meinke, M., Illenberger, E., et al., 1992, Chem. Phys., 161, 501

Tschurl, M., & Boesl, U., 2006, Int. J. Mass Spectrom., 249, 364

Turro, N.J., Ramamurthy, V., & Scaiano, J.C., 2010, “Modern Molecular Photochemistryof Organic Molecules” (University Science Books: Sausalito, CA)

Uggerud, E., 2003, Top. Curr. Chem., 225, 3

Useli-Bacchitta, F., Bonnamy, A., Mulas, G., et al., 2010, Chem. Phys., 371, 16

Vala, M., Szczepanski, J., Oomens, J., & Steill, J.D., 2009, J. Am. Chem. Soc., 131, 5784

Verstraete, L., Leger, A., d’Hendecourt, L., Dutuit, O., & Defourneau, D., 1990, Astron.Astrophys., 237, 436

Vestal, M.L., 2001, Chem. Rev., 101, 361


Yang, Z., Eichelberger, B., Carpenter, M., et al., 2010, submitted

Zhao, D., Langer, J., Oomens, J., & Dopfer, O., 2009, J. Chem. Phys., 131, 184307


[FEPAH]+ COMPLEXES AND [FEXPAHY]+ CLUSTERSIN THE INTERSTELLAR MEDIUM: STABILITY

AND SPECTROSCOPY

A. Simon1, 2, M. Rapacioli1, 2, F. Spiegelman1, 2 and C. Joblin3, 4

Abstract. The relevance of organometallic complexes in the chemistryof the interstellar medium (ISM) was first suggested in the early 90s.This initial proposal has recently been re-considered in the light ofnew astronomical data, benefiting from pioneering experimental tech-niques and theoretical methods. In this article, a review of recenttheoretical and experimental results obtained for PAHs coordinated toFe atoms is presented, focusing on: -(i)- the IR spectra of [FePAH]+

and [Fe(PAH)2]+ complexes, found to be in line with the AIBs, with

additional specific features in the far-IR range and -(ii)- the photo-evaporation of [FexPAHy]

+ clusters as possible candidates for verysmall grains releasing PAHs in photodissociation regions. These re-sults call for new astronomical data at high spatial resolution. Studieson larger clusters will require further experimental and theoretical de-velopments.

1 Introduction: The Fe-PAH proposal

Serra et al. (1992) and Chaudret et al. (1991) first proposed that π-adducts ofpolycyclic aromatic hydrocarbons (PAHs) and Fe atoms are stable species thatcan form efficiently in the conditions of the interstellar medium (ISM) and play arole in its chemistry. The efficient coordination of Fe atoms on PAHs would thenpartially account for the significant depletion of elemental Fe from the interstellargas-phase. Besides, Fe atoms or clusters trapped on carbonaceous grains could play

1 Universite de Toulouse, UPS, LCPQ (Laboratoire de Chimie et Physique Quantiques),IRSAMC, 118 Route de Narbonne, 31062 Toulouse, France2 CNRS, LCPQ, IRSAMC, 31062 Toulouse, France3 Universite de Toulouse, UPS, Centre d’Etude Spatiale des Rayonnements (CESR),Observatoire Midi-Pyrenees, 9 AV. Colonel Roche, 31028 Toulouse Cedex 04, France4 CNRS, CESR, 31028 Toulouse, France



a catalytic role in the formation of large hydrocarbons, and of PAHs in particular.This inital proposal opened a new class of species to be considered and motivatedat the time a few experimental (Marty 1996; Marty et al. 1996) and theoretical(Klotz et al. 1995) studies, along with astrophysical modeling (Marty et al. 1994;Ristorcelli & Klotz 1997).

It has recently been reconsidered in the new astrophysical context e.g.:-(i)- the search for the carriers of the blue component of the 6.2 μm aromaticinfrared band (AIB) assigned to PAH-derived species (Peeters et al. 2002) suchas substituted PAHs like PANHs+/0/− (Bauschlicher et al. 2009; Hudgins et al.2005), or PAHs coordinated to an heteroatom such as Mg (Bauschlicher 2009;Bauschlicher & Ricca 2009), Si (Joalland et al. 2010, 2009) or Fe (Simon & Joblin2007, 2010).-(ii)- the search for candidates for the VSGs located in molecular clouds and shownto release PAH molecules at the surface of the cloud (Rapacioli, see elsewhere inthis volume): the presence of heteroatoms like Fe (or Si) could stabilize PAHclusters, that are possible candidates for these VSGs (Rapacioli et al. 2005). Theexperimental study, coupled with theoretical modeling, of the photostability of[FexPAHy]+ clusters in the conditions of the ISM, indeed shows that such clusterscan be good candidates for these VSGs (Simon & Joblin 2009).

2 Stability of [FePAH]+ and [FexPAHy]+ in the ISM

Calculations based on Density Functional Theory (DFT) show that [FePAH]+

complexes are intrinsically more stable than [FePAH]0 complexes due to largerFe-PAH binding energies, e.g. ∼ 2.5 eV for [FePAH]+ complexes vs . ∼ 0.6 eV for[FePAH]0 complexes (Simon & Joblin 2007). They are thus more likely to sur-vive in the conditions of the ISM than their neutral counterparts. [FeC24H12]+

and [Fex(C24H12)2]+ (x=1,4) complexes were formed in the PIRENEA set-up,a cold ion trap dedicated to astrochemistry (Simon & Joblin 2009). Interest-ingly, complexes with dehydrogenated coronene also form when the number of Featoms reaches 3, suggesting the occurence of intramolecular organometallic reac-tions. Each [FeC24H12]+ and [Fex(C24H12)2]+ (x=1,3) was isolated and photo-dissociated under soft irradiation conditions. The major dissociation pathwayis a sequential loss of Fe atoms from [Fe3(C24H12)2]+ and [Fe2(C24H12)2]+ tolead to [Fe(C24H12)2]+ that loses one coronene molecule under further irradia-tion. The smallest complex [FeC24H12]+ then loses Fe, leaving [C24H12]+ as thefinal charged photofragment. The stochiometry of these [Fex(C24H12)2]+ com-plexes and their dissociation sequence make them good candidates for the VSGsmentioned in the previous section. The photodissociation kinetics of the smallestcomplex [FeC24H12]+ was simulated using a Monte Carlo kinetic model in orderto retrieve, from the fit of the experimental points, data such as the dissociationpre-exponential factor Ad = 1012 s−1 and the mean internal energy at dissociation(< Ud >min ∼ 5 eV). These data can now be inserted into astrochemical models.

A. Simon et al.: Clusters of Fe-Coordinated PAH in the ISM 443

3 Mid- and far-IR features of [FePAH]+ and [Fe(PAH)2]+ complexes

The AIBs are emission features from UV-excited species. Anharmonic effects aretherefore expected to play an important role in these spectra. A laboratory ap-proach to study these effects is to perform infrared multiple photon dissociation(IRMPD) spectroscopy experiments (Oomens, see elsewhere in this volume). Thistechnique was applied to [FeX1,2]+ complexes with small PAHs (X=C6H6,C10H8,C13H10 (Szczepanski et al. 2006)) and to [YFeC24H12]+ heterogeneous complexes(Y=C5H5, C5(CH3)5 (Simon et al. 2008)). To our knowledge, no IRMPD ex-periments have been performed yet on [FexPAHy]+ complexes with large PAHs,the production and dissociation of these species being challenging tasks. Infrared(IR) spectra of neutral FePAH complexes were recorded in cold rare-gas matri-ces (Elustondo et al. 1999; Wang et al. 2007). All laboratory spectra have beencompared with calculated harmonic IR spectra obtained at the DFT level and thecalculations have been extended to larger systems (Simon & Joblin 2010). Thesecalculations remain the most straightforward approach to test the contribution ofvarious PAH-derived species to the AIBs.

3.1 Harmonic spectra

Using DFT calculations, we found that the π-coordination of an Fe atom on anindividual PAH+ leads to a collapse of the band intensity in the 6.2 μm region –where the CC stretching modes (νCC) are located – with respect to the 11.2 μmregion – where the intense C-H out-of-plane bending modes (γCH) are found. Ablueshift of the νCC band and a slight redshift of the γCH band upon coordinationof an Fe atom were also put foward (Simon & Joblin 2007). Considering a mixtureof cationic PAHs ranging from pyrene C16H10 to circumcoronene C54H18 (Simon& Joblin 2010), the π-coordination of an Fe atom was shown to lead to:-(i)- an increase of the intensity ratio of the C-H stretching (νCH at 3.25±0.01 μm)and γCH bands with respect to the intense νCC band, and shifts of the νCC andγCH band positions (−0.1 μm and +0.04 μm respectively), with characteristicprofiles displaying a steep blue rise and an extended red tail,-(ii)- the occurence of many new bands in the far-IR range. Some vibrational modesof the PAH skeleton are indeed activated due to symmetry reduction and newmodes involving the motion of the Fe atom appear. In particular, an accumulationpoint due to the activation of the Fe-PAH stretching mode is observed at ∼40 μm.This range is thus suggested to contain the spectral fingerprint for the presenceof [M-PAH]+ (M=Fe, Si, Mg) complexes in the ISM. Additional features in the[60–300] μm range were also found for complexes with large PAHs.

[Fe(PAH)2]+ were shown to be stable in low-temperature and low-pressure con-ditions (Simon & Joblin 2009). The harmonic spectra of mixtures of homogeneous[Fe(PAH)2]+ and heterogeneous [(C66H20)FePAH]+ for 6 compact PAHs rangingfrom pyrene (C16H10) to circumovalene (C66H20), were computed at the DFT levelof theory (Fig. 1). Both mid-IR spectra present very similar features, making them


a prototype for [Fe(PAH)2]+ complexes with large PAHs. The 6–9 μm region dif-fers from that of [FePAH]+ complexes (Simon & Joblin 2010), with an enhancedintensity of the 7.38 μm band in particular. The intensity of the 3.25 μm band isalso considerably enhanced with respect to that of the γCH and νCC bands. In thefar-IR range, bands with reasonable intensities appear at ∼ 30 and 100 μm. Theyare assigned respectively to the Fe-PAH stretching mode coupled to PAH-skeletondeformations, and to the Fe motion parallel to the PAH surface (Simon & Joblin,in preparation).

The mid-IR spectra of [Fe(PAH)x]+ (x=1,2) are found to be good candidatesto account for both positions and profiles of the AIBs. Their rich far-IR spectracall for new astronomical data from space missions such as the Herschel SpaceObservatory and the future SPICA telescope, that can probe this spectral rangewith an increased sensitivity.

0

5000

10000

15000

20000

2 4 6 8 10 12 14

I (km/mol/μm)

wavelength (μm)

[(C66

H20

)FePAH]+

[Fe(PAH)2]+

3.25

6.18

7.38 11.1211.15

Fig. 1. Left: computed mid-IR harmonic spectra of mixtures of [Fe(PAH)2]+ (red) and

[(C66H20)FePAH]+ (blue) complexes. Right: stable structure of [Fe(C66H20)2]+ obtained

at the DFT level of theory.

3.2 Anharmonic effects on the mid-IR spectra of [FePAH]+

The anharmonic effects in the mid-IR spectra of various PAH, PAH+ and [SiPAH]+

complexes were obtained from classical molecular dynamics (MD) simulations(Joalland et al. 2010) on a potential energy surface described at the DFTB (Den-sity Functional Tight Binding) level of theory (Elstner et al. 1998). Using a similarapproach, we performed a comparative study of the anharmonic effects on the IRspectra of [FeC24H12]+ and [C24H12]+ (Simon et al. submitted), showing in par-ticular that:-(i)- the anharmonicity of the γCH band is decreased upon coordination of an Featom, resulting in the merging of the γCH bands of [FeC24H12]+ and [C24H12]+

at ∼ 700 K (Fig. 2),

A. Simon et al.: Clusters of Fe-Coordinated PAH in the ISM 445

-(ii)- the anharmonicity of the νCC band is also decreased, enhancing the blueshiftof the νCC band of [FeC24H12]+ with respect to that of [C24H12]+.The band positions were shown to have a linear dependence on temperature. Ex-trapolating these dependences to the mixture of [FePAH]+ complexes with largePAHs studied by Simon & Joblin (2010), we found that the γCH band shifts from11.02 μm at 0 K to 11.23 μm at 800 K whereas the νCC band shifts from 6.16 μmto 6.31 μm. These results show that the assignment of the 6.2 vs . 6.3 μm compo-nents of the AIBs should be done with caution.MD simulations also show that the diffusion of the Fe atom on the [C24H12]+ sur-face is observed at T ∼ 850 K, e.g. for an energy excess of ∼ 4 eV, which is belowthe dissociation energy determined to be 5 eV (cf. Sect. 2). This DFTB/MDmethod also allows us to probe high-temperature chemistry and will be used forreactivity studies in a very near future.

0

1

2

3

4

5

6

7

830 840 850 860 870 880 890

arbitraryunit

wavenumber (cm-1)

C24

H12

[C24

H12

]]+[FeC24

H12

]+

550K

350K

150K

50K

650K

750K

Fig. 2. Left: DFTB/MD simulated anharmonic IR spectra (zoom on the CH out-of-plane

bending mode γCH) of [FeC24H12]+ (red) and [C24H12]

0/+ (green/black). Right: most

stable structure of [FeC24H12]+ obtained at the DFT level of theory.

4 Conclusions and future work

The computed harmonic mid-IR spectra of [FePAH]+ and [Fe(PAH)2]+ complexeswith large PAHs show that these species could contribute to the AIBs. The far-IRspectra of these complexes present specific features that need to be searched for inastronomical spectra. Experimental results show that [FexPAHy]+ complexes arelikely candidates for astronomical VSGs releasing PAHs at the surface of molecularclouds. These studies have benefited from recent experimental and theoreticaldevelopments that will be pursued to improve our knowledge of larger clusters.Future work will include the study of the stability and spectroscopy of complexesof Fe atoms with dehydrogenated PAHs and the investigation of the role of FePAHcomplexes in the formation of H2.


References

Bauschlicher, C.J., 2009, Mol. Phys., 107, 809

Bauschlicher, C.W., Peeters, E., & Allamandola, L.J., 2009, Astrophys. J., 697, 311

Bauschlicher, C.W., Jr., & Ricca, A., 2009, Astrophys. J., 698, 275

Chaudret, B., Le Beuze, A., Rabaa, H., Saillard, Y., & Serra, G., 1991, New J. Chem.,15, 791

Elstner, M., Porezag, D., Jungnickel, G., et al., 1998, Phys. Rev. B, 58, 11, 7260

Elustondo, F., Dalibart, M., Derouault, J., & Mascetti, J., 1999, Phys. Chem. Earth C,24, 583

Hudgins, D.M., Bauschlicher, C.W., & Allamandola, L.J., 2005, Astrophys. J., 632, 1,316

Joalland, B., Rapacioli, M., Simon, A., et al., 2010, J. Phys. Chem. A, 114, 5846

Joalland, B., Simon, A., Marsden, C.J., & Joblin, C., 2009, Astron. Astrophys., 494, 969

Klotz, A., Marty, P., Boissel, P., et al., 1995, Astron. Astrophys., 304, 520

Marty, P., 1996, Chem. Phys. Lett., 256, 669

Marty, P., de Parseval, P., Klotz, A., Serra, G., & Boissel, P., 1996, Astron. Astrophys.,316, 270

Marty, P., Serra, G., Chaudret, B., & Ristorcelli, I., 1994, Astron. Astrophys., 282, 916

Peeters, E., Hony, S., Van Kerckhoven, C., et al., 2002, Astron. Astrophys., 390, 1089

Rapacioli, M., Joblin, C., & Boissel, P., 2005, Astron. Astrophys., 429, 193

Ristorcelli, I., & Klotz, A., 1997, Astron. Astrophys., 317, 962

Serra, G., Chaudret, B., Saillard, Y., et al., 1992, Astron. Astrophys., 260, 489



Simon, A., & Joblin, C., 2010, Astrophys. J., 712, 69

Simon, A., & Joblin, C., in preparation

Simon, A., Joblin, C., Polfer, N., & Oomens, J., 2008, J. Phys. Chem. A, 112, 8551

Simon, A., Rapacioli, M., Lanza, M., Joalland, B., & Spiegelman, F., Phys. Chem. Chem.Phys., submitted

Szczepanski, J., Wang, H., Vala, M., et al., 2006, Astrophys. J., 646, 666

Wang, Y., Szczepanski, J., & Vala, M., 2007, Chem. Phys., 342, 107


MODELLING THE PHYSICAL AND CHEMICAL EVOLUTIONOF PAHS AND PAH-RELATED SPECIES IN

ASTROPHYSICAL ENVIRONMENTS

J. Montillaud1, 2, C. Joblin1, 2 and D. Toublanc1, 2

Abstract. An active carbon chemistry is observed at the border ofphoto-dissociation regions (PDRs), involving small hydrocarbons, poly-cyclic aromatic hydrocarbon (PAH) macromolecules and evaporatingvery small grains (VSGs). In this context, we aim at quantifying thephysical and chemical evolution of PAHs (hydrogenation and chargestates, aggregation, and complexation with heavy atoms) as a functionof the local physical conditions (radiation field, temperature, density,abundances of atomic and molecular hydrogen, electrons and heavyatoms). We have developed a numerical model that follows the timedependency of the abundance and internal energy of each species. Inthis paper, we use this model to calculate the hydrogenation and chargestates of coronene C24H12 as an interstellar PAH prototype. We takeadvantage of recent results on photodissociation and reaction ratesand provide guidelines for future laboratory studies. Reaction ratesof coronene-derived radical cations with H and H2 are found to besufficiently constrained by experiments, whereas the absence of exper-imental data for neutral species is critical.

1 Introduction

During the two last decades, polycyclic aromatic hydrocarbons (PAHs) have beenextensively studied as candidate carriers for the so-called aromatic infrared bands(AIBs). In order to test this hypothesis, models of increasing complexity andaccuracy have been developed. Bakes & Tielens (1994) described the charge stateof PAHs and their contribution to the photoelectric heating. Allain et al. (1996a,1996b) modelled the photodissociation rates of PAHs, taking into account the lossof H, H2, and C2H2, and discussed their lifetime in the interstellar medium (ISM).

1 Universite de Toulouse, UPS, CESR, 9 Av. du Colonel Roche, 31028 Toulouse Cedex 4,France2 CNRS, UMR5187, 31028 Toulouse, France



Le Page et al. (2001, 2003) further improved the description of the photophysicaland chemical processes of PAHs and discussed the hydrogenation and charge statesof PAHs in the diffuse ISM. Visser et al. (2007) built a model based on previousstudies to describe the evolution of PAHs in protoplanetary disks.

The accuracy of the results provided by these models depends on how well pho-tophysical and chemical processes can be quantified. Experimental data are oftenscarce and studies on fundamental processes are needed, which in general involveinterdisciplinary collaborations between physicists, chemists, and astrophysicists.

In this paper, we analyse the question of the charge and hydrogenation states ofcoronene (C24H12) taking advantage of experimental data obtained with the coldion trap PIRENEA (Joblin et al. 2002a). Section 2 presents the model that wasdeveloped to perform this study. The chemical and physical data are discussed inSection 3. Results are given and discussed in Section 4.

2 Presentation of the model

The model includes all hydrogenation states from fully dehydrogenated to super-hydrogenated coronene (0 to 24 H). We do not consider PAH anions since there isso far no clear evidence for their presence in the ISM and we are facing a lack ofinformation in particular on their reactivity with hydrogen. The processes drivingthe evolution of charge and hydrogenation states are discussed in Section 3.

The time evolution of the species is computed using the rate equation for-malism. Typical environments are PDRs in which the physics and chemistry aredriven by UV photons. In order to describe as properly as possible the physicalconditions in these PDRs, we use the Meudon PDR code (Le Petit et al. 2006)that provide a self-consistent set of physical parameters, as a function of depthinside the cloud.

3 Charge and hydrogenation states

3.1 Charge balance

In the scope of this work, we focus on the most external layers of PDRs. ThePAH charge state is determined by the balance between the photoionization andthe electron recombination with PAH cations. The photoionization rate for agiven energy of the UV photon hν is computed as the product of the absorptioncross-section σabs(hν) by the photon flux Fphoton(hν) and by the ionization yieldYion(hν). The values of σabs are from Malloci et al. (2004) and those of Yion fromVerstraete et al. (1990).

Electron recombination rates have been measured for a few small PAH cations,up to the pyrene cation C16H+

10 (see Biennier et al. 2006). The classical law fora thin conducting disk (e.g. Verstraete et al. 1990) overestimates the rates forthese small species, but is expected to be relevant for large PAH cations (NC ≈100). For species of intermediate size like coronene cation C24H+

12, we assumethe recombination rate to be between the value of pyrene (4.1 × 10−6 cm3 s−1 at

J. Montillaud et al.: Modelling the Evolution of PAHs 449

300K) and the value predicted by the classical law (1.6× 10−5 cm3 s−1 at 300K).We choose the mean value krec = 1.0 × 10−5 cm3 s−1 at 300K and also considerthe extreme cases krec = 4.1 × 10−6 cm3 s−1 and krec = 1.6 × 10−5 cm3 s−1. Adependence of krec on T−1/2 is used (Bakes & Tielens 1994).

For the sake of simplicity, the photoionization and electron recombination rateswere supposed not to depend on the hydrogenation state.

3.2 Reactivity of PAHs with hydrogen

Only few studies have been dedicated to the reactivity of PAH cations with Hor H2, and we found no data for the reactivity of neutral PAHs. We use notemperature dependence for neutral-ion reactions, and use reaction rates k ∝ T 1/2

for neutral-neutral reactions. All the values hereafter are given at 300 K exceptwhen mentioned.

The reactivity of PAH cations with H has been studied experimentally fora few PAH cations (Bierbaum, see elsewhere in this volume) including coronene(k+H = 1.4 ± 0.7 × 10−10 cm3 s−1, Betts et al. 2006). We generalized the lattervalue to all coronene derivative species missing an even number of hydrogens, i.e.radical dehydrogenated cations. Cations missing an odd number of hydrogens areexpected to be less reactive. This was demonstrated on naphthalene by Le Pageet al. (1997), who measured k+H = 1.9 × 10−10 and < 5 × 10−11 cm3 s−1 forC10H+

8 and C10H+7 , respectively. Since no experimental data are available for

closed-shell dehydrogenated coronene cations, we follow Le Page et al. (2001) anduse a reaction rate of k+H = 5 × 10−11 cm3 s−1.

Betts et al. (2006) have evaluated an upper limit for the reactivity of C24H+12

with H2 of 5 × 10−13 cm3 s−1. We generalize this result to all radical dehydro-genated cations. Similarly, recent measurements using the cold ion trap PIRENEA(Joblin et al. 2002a) have shown no reactivity of C24H+

11 with H2 at low pressure(∼ 10−8 mbar ) and temperature (T <≈ 80 K). An upper limit of 4×10−12 cm3 s−1

was derived. Here again, we generalize this value to all closed-shell dehydrogenatedcoronene cations.

Superhydrogenated naphthalene and pyrene were measured not to react withH2, and to react slowly with H at rates around ∼10−12 cm3 s−1 (Snow et al. 1998).Hence we assume superhydrogenated coronene to react with H only, at a rate of10−12 cm3 s−1. More informations concerning superhydrogenated PAHs can befound in Thrower et al. (see elsewhere in this volume).

3.3 Photodissociation of coronene by hydrogen loss

After absorption of a UV photon, a PAH can either dissociate or cool down radia-tively. Allain et al. (1996a) showed that the relaxation cascade is dominated byIR photons. The IR emission rates were calculated in a microcanonical formalism(Joblin et al. 2002b).

Using the PIRENEA set-up, Joblin et al. studied the photodissociation ofC24H+

p species (p = [1, 12]) and derived dissociation rates as a function of the


Table 1. Summary of the reaction rates (cm3 s−1) at 300 K used in our model for coronene

derivatives and their origin. The neutral counterparts are assumed to react neither with

H nor with H2. (a) Betts et al. (2006); (b) Le Page et al. (2001); (c) Snow et al. (1998);

(d) No detection with the PIRENEA set-up.

Species C24H+n (even n ≤ 12) C24H

+n (odd n < 12) C24H

+n (n > 12)

+H 1.4 × 10−10 (a) 5.0 × 10−11 (b) 1.0 × 10−12 (c)+H2 < 5 × 10−13 (a) < 4 × 10−12 (d) - (b)+e− 1.0 × 10−5 (extrapolation)

1e-06

0.0001

0.01

1

100

10000

1e+06

4 6 8 10 12

k(s-1

)

E(eV)

kIR from Le Page et al. (2001)

kIR from this work

kdiss from Le Page et al. (2001)kdiss from this work

Fig. 1. Comparison of the IR cooling (dashed lines) and photodissociation (solid lines)

rates from Le Page et al. (2001) (green lines) and from this work (red lines) for the

coronene cation C24H+12, as a function of its internal energy. Vertical lines indicate the

threshold energy above which the dissociation is faster than the IR emission.

internal energy (Joblin et al. 2010a). Only H loss was observed for these species,the destruction of the carbon skeleton occuring only after complete dehydrogena-tion (C+

24; Joblin et al. 2003; Joblin et al. 2010b). The reaction rates derivedby Joblin et al. differ significantly from the values from Le Page et al. (2001; cf.Fig. 1).

4 Results

Using our numerical model, we compute the hydrogenation and charge states ofcoronene for the northern PDR in NGC 7023. For the incident radiation field, weused the stellar spectrum from the Kurucz library (Kurucz 1991) for a star witha temperature of 15 000 K and an integrated UV intensity of 2600 Habing. Thedensity was set to nH = 2 × 104 cm−3 (Gerin et al. 1998).

General behaviour. The hydrogenation state of coronene adopts four possi-ble regimes: (1) totally dehydrogenated (C0/+

24 dominates), (2) partially dehy-drogenated with a roughly equivalent abundance for each C24H+

p , species (p =

[1, 11]), (3) normally hydrogenated (C24H0/+12 dominates) and (4) superhydro-

genated (C24H0/+13 dominates).

Hydrogenation and charge state of coronene in NGC 7023. As shown inFigure 2, coronene is found to be essentially neutral throughout the PDR. Despite

J. Montillaud et al.: Modelling the Evolution of PAHs 451

the high intensity of the radiation field, PAHs are ionized only in the first layers ofthe cloud. The totally dehydrogenated species, C24, dominates inside the PDR upto a depth of AV ≈ 5, where G0 ≈ 20 in Habing units. For equivalent conditions,i.e. G0/nH ≈ 1/1000, Le Page et al. found the normal hydrogen coverage to dom-inate. This difference is consistent with our model using larger photodissociationrates, and lower reaction rates with H2.

0

0.2

0.4

0.6

0.8

1

0 20 40 60 80 100 120

0 2.5 5 7.5 10 12.5 15

Rel

ativ

e w

eigh

t

Position (arcsec)

Position (Av)

neutrals

cations

C240/+

C24H12-n0/+

C24H120/+C24H13

0/+

C240/+

C24H120/+

C24H12-n0/+

Fig. 2. Evolution of the charge and hydrogenation states of coronene through the north-

ern PDR of the reflection nebula NGC 7023. Black lines: evolution of the abundances

of cations (dashed lines) and neutrals (solid lines), for different electron recombination

rates (thick lines: krec(300 K) = 1.0 × 10−5 cm3 s−1; thin lines: krec(300 K) = 0.4 and

1.6 × 10−5 cm3 s−1). Color lines: abundances of C0/+24 (dashed lines), partially dehy-

drogenated species C24H+12−n with n=[1,11] (dot-dashed lines), C24H

+12 (solid lines) and

C24H+13 (dotted lines) derived with (i) our standard model (red), (ii) maximum values

for reactivity between cations and H2 (blue), (iii) reaction rates of neutral radicals with

H set to 1.4 × 10−11 cm3 s−1 at 300 K (orange) and (iv) reaction rates of neutrals with

H set at 300 K to 1.4 × 10−11 cm3 s−1 (radicals) and 5.0 × 10−12 cm3 s−1 (closed-shells)

(green).

Influence of the uncertainties. The uncertainty on the electron recombinationrate affects the ionization fraction to at most a factor of 2. Tuning the reactionrates of PAH cations with H2 from their upper limit to zero does not changesignificantly the results. The abundances are also not strongly affected by the ex-perimental uncertainty of 50% on the reactivity of C24H+

12 with H. On the contrary,tuning the reactivity of neutral PAHs, which was set to 0 in our standard model,leads to substantial changes. If neutral radicals C24H2n−1 react with H at a rateat 300K ten times lower than cation radicals, the hydrogenation state differs onlydeep in the cloud (AV > 10), where PAHs are very likely in solid phase. However,if neutral closed-shell species C24H2n are also considered to react with H, with arate at 300K ten times lower than for closed-shell cations, then C24H12 dominatesbetween AV = 6 and 9, while C24H13 dominates deeper.


5 Conclusion

We present the first results of a model dedicated to the study of the evolutionof PAHs in astrophysical environments. We examine the case of coronene hydro-genation and charge states, in the northern PDR of NGC 7023, using the latestavailable physical and chemical data. Coronene was found to be much more dehy-drogenated than calculated in previous studies. Reaction rates of coronene-derivedradical cations with H and H2 were found to be sufficiently constrained by exper-iments, whereas the absence of experimental data for the reactivity of neutralspecies appears to be critical. A first theoretical study is presented elsewhere inthis volume (Thrower et al.).

References

Allain, T., Leach, S., & Sedlmayr, E., 1996a, A&A, 305, 602

Allain, T., Leach, S., & Sedlmayr, E., 1996b, A&A, 305, 616



Biennier, L., Alsayed-Ali, M., Foutel-Richard, A., et al., 2006, Faraday Discussions, 133,289

Draine, B.T., & Sutin, B., 1987, ApJ, 320, 803

Gerin, M., Phillips, T., Keene, J., Betz, A.L., & Boreiko, R.T., 1998, ApJ, 500, 329

Joblin, C., Pech, C., Armengaud, M., Frabel, P., & Boissel, P., 2002a, EAS PublicationSeries, 4, 73

Joblin, C., Toublanc, D., Boissel, P., & Tielens, A.G.G.M., 2002b, Mol. Phys., 100, 3595

Joblin, C., 2003, in SF2A-2003, ed. F. Combes et al. (EDP-Sciences), 175

Joblin C., Toublanc D., Pech C., Armengaud M., Frabel P., & Boissel P., 2010a, inpreparation

Joblin C., Armengaud M., Frabel P., & Boissel P., 2010b, in preparation

Jochims, H.W., Ruhl, E., Baumgartel, H., Tobita, S., & Leach, S., 1994, ApJ, 420, 307

Kurucz, R.L., 1991, BAAS, 23, 1047

Leger, A., d’Hendecourt, L., Boissel, P., & Desert, F.X., 1989, A&A, 213, 351

Le Page, V., Keheyan, Y., Bierbaum, V.M., & Snow, T.P., 1997, JACS, 119, 35

Le Page, V., Keheyan, Y., Snow, T.P., & Bierbaum, V.M., 1999a, IJMS, 185, 949

Le Page, V., Keheyan, Y., Snow, T.P., & Bierbaum, V.M., 1999b, JACS, 121, 9435









SUPERHYDROGENATED PAHS: CATALYTIC FORMATIONOF H2

J.D. Thrower1, L. Nilsson1, B. Jørgensen1, S. Baouche1, R. Balog1,A.C. Luntz1, I. Stensgaard1, E. Rauls2 and L. Hornekær1

Abstract. The possible role of neutral PAHs as catalysts for H2 forma-tion in the interstellar medium is investigated by a combined exper-imental and density function theory study of the superhydrogenationof coronene (C24H12). The calculations suggest efficient hydrogenationof both edge and centre sites, along with competing abstraction reac-tions to form H2 in a series of catalytic cycles. Scanning tunnelingmicroscopy and thermal desorption measurements have been used toprovide direct evidence of the formation of superhydrogenated coroneneas a result of exposure to D atoms. Lower limit estimates for the cross-sections of 1.8×10−17 , 5.5×10−18 and 1.1×10−18 cm2 for the formationof singly, doubly and triply hydrogenated coronene are derived. Theresults suggest that superhydrogenated PAHs may play an importantrole in H2 formation in the ISM.

1 Introduction

The interstellar formation of H2 remains the subject of intense research effort giventhe importance of this molecule as a cooling agent and as a key precursor for chem-ical processes in interstellar dust and molecular clouds. The present consensus isthat molecular hydrogen forms on the surface of interstellar dust grains by reac-tions between hydrogen atoms (Hollenbach & Salpeter 1971). Experiments andtheoretical calculations have shown that surface reactions involving physisorbedH atoms are efficient at temperatures below approximately 20 K (Pirronello et al.1997b; Pirronello et al. 1997a; Pirronello et al. 1999; Katz et al. 1999; Manicoet al. 2001; Roser et al. 2002; Hornekær et al. 2003; Cuppen & Herbst 2005).

1 Department of Physics and Astronomy, Aarhus University, 8000 Aarhus C, Denmarke-mail: [email protected] Department of Theoretical Physics, Faculty of Natural Sciences, University of Paderborn,33098 Paderborn, Germany



In regions with gas temperatures of several hundred kelvin surface reactions in-volving chemisorbed H atoms have been shown to be efficient (Zecho et al. 2002;Cazaux & Tielens 2004; Hornekær et al. 2006a, 2006b). However, at intermediatetemperatures, no efficient routes for H2 formation have been identified (Cuppen& Herbst 2005).

A possible alternative route to interstellar H2 formation is provided by PAHmolecules. Observations of spatial correlations between vibrationally excited H2

and PAHs in photo dissociation regions provide indications of PAHs acting as cat-alysts for H2 formation (Habart et al. 2003; Habart et al. 2004). PAHs in the ISMare thought to exhibit a range of hydrogenation and charge states depending onthe environment, especially the UV flux (Le Page et al. 2001; Le Page et al. 2003;Montillaud et al. 2011). At high UV fluxes small PAHs will be dissociated, whilstat intermediate fluxes H loss and ionization dominate, resulting in dehydrogenatedand cationic PAHs. Under low UV flux conditions neutral, anionic and superhy-drogenated PAHs are likely to be present in the gas phase or frozen out on dustgrain surfaces. For example, observations of the 3.4 μm emission aliphatic band inlow UV flux regions compare well with laboratory spectra of superhydrogenatedneutral PAHs (Bernstein et al. 1996). To date, investigations of H2 formationon PAHs have focussed primarily on cationic states, see, for example, (Le Pageet al. 1997; Bauschlicher 1998; Betts et al. 2006; Le Page et al. 2009; Bierbaum2011). However, formation of H2 on PAH anions (Bauschlicher & Bakes 2001) andneutral PAHs (Stein & Brown 1991; Bauschlicher 1998; Rauls & Hornekær 2008;Rodriguez et al. 2010), as well as on PAH coated grains (Duley & Williams 1993)has also been discussed.

Here we investigate in detail the possible role of neutral PAH molecules andPAH covered dust grains as catalysts for H2 formation. Density functional the-ory (DFT) calculations on the hydrogenation of the coronene molecule reveal arelatively small barrier to the addition of a single H atom to coronene. Interest-ingly, subsequent hydrogenation and abstraction reactions can proceed with littleor no barrier, suggesting a catalytic cycle which results in H2 formation (Rauls& Hornekær 2008). Experimental observations of superhydrogenated coroneneformed through exposure to atomic D are also presented. These combined ex-perimental and theoretical results suggest that superhydrogenation of PAHs mayprovide an important additional route to H2 formation in intermediate, as well ashigh temperature environments in the ISM. Similar experimental results for thecase of aromatic carbon grain mimics are provided by Mennella (2011).

2 Methods

The DFT calculations were carried out using the plane wave based DACAPOcode (Hammer et al. 1999; Bahn & Jacobsen 2002) and the PW91 exchangecorrelation (xc) functional (Perdew et al. 1992). Spin polarization was includedin all calculations. Potential energy curves were constructed by performing single

J.D. Thrower et al.: Superhydrogenated PAHs 455

Outer edge

Edge Centre

Fig. 1. The possible hydrogenation sites on the coronene molecule. The sites are referred

to as outer edge (oe), edge (e) and centre (c).

point calculations with constrained distances between either one H and one Catom, or two H atoms in the case of hydrogenation and abstraction reactionsrespectively. Barrier heights were obtained by iterative energy calculations withinclose proximity of the critical geometry. As a result, all barriers can be consideredto be upper bounds for the activation energies. See Rauls & Hornekær (2008) forfurther details.

Scanning tunneling microscopy (STM) and thermal desorption measurementswere performed under ultrahigh vacuum (UHV) conditions in two separate cham-bers. In both chambers, C24H12 films were grown by placing the substrate close toan evaporation source in which the C24H12 sample was held at 180◦C. The coronenefilms were then exposed to atomic D, which was produced using a hot capillarythermal cracker source of the Julich type (Tschersich & von Bonin 1998). Tem-perature programmed desorption (TPD) experiments employed a highly orientedpyrolytic graphite (HOPG) substrate, cleaved prior to mounting in the UHV cham-ber. TPD spectra were acquired using a quadrupole mass spectrometer (QMS)and a heating ramp of 1 K s−1. A Cu(100) single crystal substrate was used forthe STM measurements to exploit the improved imaging conditions provided bythe stronger binding between coronene molecules and this substrate.

3 Results

There are three possible sites for the addition of a single H atom to coronene asshown in Figure 1. The most stable product is obtained through addition to acarbon atom in an outer edge site. The additional H atom has a binding energyin excess of 1.4 eV, twice that found for graphite. There is a barrier to addition of60 meV which is significantly smaller the 200 meV found for H adsorption on thebasal plane of graphite (Sha & Jackson 2002). This indicates that hydrogenationof C24H12 is likely to occur in cooler environments than for graphite and that thehydrogenated species thus formed will be stable.


Fig. 2. A summary of the results of the DFT calculations. The central line shows the

most favourable addition reactions. The sites involved are indicated. The abstraction

reactions and catalytic loops are shown on the left. All energies are the associated reaction

barriers. The top and bottom sides of the molecule are labelled t and b respectively.

Reproduced by permission of the AAS from Rauls & Hornekær (2008).

Further addition reactions were also considered. The results of these DFT cal-culations for up to 8 additional H atoms, as well as abstraction reactions from eachhydrogenation state, are summarized in Figure 2. Individual potential energy sur-faces for these reactions, along a more detailed discussion, can be found elsewhere(Rauls & Hornekær 2008). Addition of a second H atom is most favourable tothe outer edge site adjacent to that involved in the first hydrogenation step. Thisreaction is barrierless and results in a binding energy for the second H atom of3.2 eV which gives a combined binding energy for the outer edge dimer of ca.4.7 eV. Comparing this value with the H2 binding energy of 4.5 eV indicates thatdirect recombination of two bound H atoms to form gas phase H2 is not energet-ically favourable. H2 formation is likely to occur, however, through Eley-Ridealabstraction reactions. Indeed, abstraction by, rather than addition of, the secondH atom is also found to be barrierless revealing a competition between these twoprocesses. Similarly, for subsequent incoming H atoms there exist reaction path-ways for both addition and abstraction that are thermodynamically viable andhave small or even vanishing activation barriers.

Interestingly, not only outer edge and edge sites can be accessed, but a favou-rable addition to a centre site is also possible with the addition of a sixth H atom.


Fig. 3. STM images (150×150 A2) of a monolayer of coronene adsorbed on Cu(100)

(a) prior to and (b) after exposure to D atoms (4×1013 cm−2). The bright protrusions of

diameter ca. 11.5 A are identified as individual coronene molecules. Imaging parameters:

(a) Vt = −2460 mV, It = −0.21 nA (b) Vt = −784 mV, It = −0.17 nA.

Addition to centre sites occurs such that adjacent H atoms are bound on oppositefaces of the molecule. This is consistent with the molecular structure approach-ing that of the fully hydrogenated coronene molecule, perhydrocoronene (C24H36)which exhibits this arrangement of hydrogen atoms. It is clear that the mostimportant step in the hydrogenation process is the addition of the first hydro-gen atom. This opens up a series of reaction pathways in which further H atomsmay be added with H2 formation occuring via catalytic loops, re-forming lowerhydrogenated states. The branching ratios for the competing addition and hydro-genation reactions are not known and will be essential for a full understanding ofthe efficiency of H2 formation.

In Figure 3a an STM image of a self-assembled monolayer of C24H12 depositedon the clean Cu(100) surface is displayed. An ordered structure of circular protru-sions of diameter ca. 11.5 A is observed. These bright protrusions are identifiedas individual coronene molecules adsorbed flat on the surface. A single vacancydefect is observed in the top part of the image. The coronene covered surface wassubsequently exposed to D atoms with a fluence of 4 × 1013 cm−2. An STM im-age of the surface after exposure is shown in Figure 3b. The image clearly showsvariations in imaging intensity between coronene molecules, as well as within in-dividual coronene molecules. These variations are attributed to the formation ofhydrogenated coronene. The range of different intensity variations in the imagesuggests the presence of coronene molecules with differing degrees of hydrogena-tion. Comparison between STM images and theoretical calculations should makeit possible to determine the number and position of deuterium atoms added toindividual coronene molecules in the near future.

Thermal desorption measurements have also been conducted using a layer ofC24H12 grown on the HOPG substrate and subsequently exposed to D atoms with afluence of 9×1015 cm−2. The substrate was then heated linearly and the desorbing


Table 1. Contributions to the desorption signal from the different hydrogenation states

along with estimated lower limits for the associated addition reaction cross-sections.

m/z % of desorption signal cross-section / cm2

300 77 —302 16 1.8 × 10−17

304 5 5.5 × 10−18

306 1 1.1 × 10−18

species detected. Desorption signals (not shown) for m/z = 300, 302, 304 and 306corresponding to coronene and singly, doubly and triply hydrogenated (with D)products were observed. Table 1 shows the relative contributions to the totaldesorption signal for the individual hydrogenation states observed. The estimatedreaction cross-sections are based on the following assumptions: only contributionfrom one coronene layer (correct within a factor of 2–3), no saturation effects,detection of all hydrogenated species. For the doubly and triply hydrogenatedcoronene the values derived are the effective overall cross-sections obtained usingthe total initial coronene coverage. Since the electron impact ionization employedin the QMS is likely to lead to fragmentation and dehydrogenation, the calculatedcross-sections are only lower limits. The cross-sections thus derived are largerthan those found previously for the irradiation of carbon grain mimics (ca. 2 ×10−18 cm2) with H atoms at temperatures of 80–300 K (Mennella et al. 2002;Mennella 2006). This is consistent with the use of significantly higher D atomtemperatures in the present study. At higher D atom fluences, even higher degreesof superhydrogenation of coronene are observed.

4 Conclusions

DFT calculations indicate the presence of energetically favourable pathways to H2

formation through catalytic cycles of hydrogenation and abstraction. The barrierto addition of the first H atom is of the order of 60 meV and small or vanishingbarriers are found for all subsequent reactions. The hydrogenation of coronenehas been observed directly using STM through the appearance of sub-molecularstructure following exposure to D atoms. There is evidence for the presence ofseveral different hydrogenated species, although comparison with theory will berequired to determine the degree of hydrogenation and the sites involved. Thermaldesorption measurements show that very high levels of superhydrogenation can beattained. Hence, both theoretical calculations and experimental results indicatethat PAHs might well play a role as catalysts for H2 formation in intermediateand high temperature environments.


The authors gratefully acknowledge financial support from the European Research Council underERC starting grant HPAH, No. 208344. The research leading to these results has receivedfunding from the European Community’s Seventh Framework Programme under grant agreementNo. 238258.

References

Bahn, S.R., & Jacobsen, K.W., 2002, Comput. Sci., 4, 56

Bauschlicher, C.W., 1998, ApJ, 509, L125

Bauschlicher, C.W., & Bakes, E.L.O., 2001, Chem. Phys., 274, 11

Bernstein, M.P., Sandford, S.A., & Allamandola, L.J., 1996, ApJ, 472, L127


Bierbaum, V.M., 2011, this volume

Cazaux, S., & Tielens, A.G.G.M., 2004, ApJ, 604, 222

Cuppen, H.M., & Herbst, E., 2005, MNRAS, 361, 565

Duley, W.W., & Williams, D.A., 1993, MNRAS, 260, 37

Habart, E., Boulanger, F., Verstraete, L., et al., 2003, A&A, 397, 623

Habart, E., Boulanger, F., Verstraete, L., Walmsley, C.M., & des Forets, G.P., 2004,A&A, 414, 531

Hammer, B., Hansen, L.B., & Norskov, J.K., 1999, Phys. Rev. B, 59, 7413

Hollenbach, D., & Salpeter, E.E., 1971, ApJ, 163, 155

Hornekær, L., Baurichter, A., Petrunin, V.V., Field, D., & Luntz, A.C., 2003, Science,302, 1943

Hornekær, L., Rauls, E., Xu, W., et al., 2006a, Phys. Rev. Lett., 97, 186102

Hornekær, L., Sljivancanin, Z., Xu, W., et al., 2006b, Phys. Rev. Lett., 96, 156104

Katz, N., Furman, I., Biham, O., Pirronello, V., & Vidali, G., 1999, ApJ, 522, 305




LePage, V., Keheyan, Y., Bierbaum, V.M., & Snow, T.P., 1997, J. Am. Chem. Soc., 119,8373

Manico, G., Raguni, G., Pirronello, V., Roser, J.E., & Vidali, G., 2001, ApJ, 548, L253

Mennella, V., 2006, ApJ, 647, L49


Mennella, V., 2011, this volume

Montillaud, J., Joblin, C., Toublanc, D., et al., 2011, this volume

Perdew, J.P., Chevary, J.A., Vosko, S.H., et al., 1992, Phys. Rev. B, 46, 6671

Pirronello, V., Biham, O., Liu, C., Shen, L.O., & Vidali, G., 1997a, ApJ, 483, L131

Pirronello, V., Liu, C., Roser, J.E., & Vidali, G., 1999, A&A, 344, 681

Pirronello, V., Liu, C., Shen, L., & Vidali, G., 1997b, ApJ, 475, L69

Rauls, E., & Hornekær, L., 2008, ApJ, 679, 531


Rodriguez, L.S., Ruette, F., Sanchez, M., & Mendoza, C., 2010, J. Mol. Catal. A, 316,16

Roser, J.E., Manico, G., Pirronello, V., & Vidali, G., 2002, ApJ, 581, 276

Sha, X.W., & Jackson, B., 2002, Surf. Sci., 496, 318

Stein, S.E., & Brown, R.L., 1991, J. Am. Chem. Soc., 113, 787

Tschersich, K.G., & von Bonin, V., 1998, J. Appl. Phys., 84, 4065

Zecho, T., Guttler, A., Sha, X., Jackson, B., & Kuppers, J., 2002, J. Chem. Phys., 117,8486

Summary of the Meeting


SUMMARY OF THE MEETING

A. Omont1

Abstract. This symposium has shown that the field of astrophysicalPAHs remains extremely active and lively. Thanks to Spitzer SpaceTelescope, the number of PAH papers has spectacularly increased, in-cluding now up to the young Universe. Laboratory and theoreticalworks have progressed in proportion. Salient features of the six ses-sions of the symposium are briefly reviewed. Comprehensive analy-ses of the rich and complex infrared spectra of interstellar PAHs arenow well established, based on a large database of observational data.PAHs are fully confirmed as excellent tracers of star formation, buttheir emission strongly depends on metallicity. Various observations,especially in harsh environments, have confirmed the complexity of thelifecycle of PAHs in space, and the need for multiple formation modes.Electronic properties remain a major issue for astronomical PAHs, in-cluding their possible connection with the diffuse interstellar bands,and the possible importance of protonated PAHs. Progress in studyingcomplex carbonaceous compounds, such as those of various soots, andin synthesizing very large PAHs may give important clues for under-standing interstellar PAHs. Significant progress was also reported inmodeling the important role of PAHs in the physics and chemistry ofthe interstellar medium.

1 Introduction

Coming back to this field after some time, I am not surprised to find it quitemature for its 25th anniversary. However, I am also struck seeing it so active andlively. Of course, as any well established field, after the explosion of its opening,it had to face renewal and serious challenges. But this symposium has shown thatnew directions are constantly opened, while older ones are renewed and benefitfrom breakthroughs in observational, laboratory and theoretical capabilities. Thesize of the attendance of this meeting including a very lively, enthusiastic and

1 Institut d’Astrophysique de Paris, UMR 7095, CNRS, UPMC Univ. Paris 06, 98bis Bd.Arago, 75014 Paris, France



outstanding new generation is the best proof of the current impact of the field andits appeal to young researchers.

Before spotting various highlights of the meeting, I would like to stress twomajor features that we may agree on being the main drivers of many of the newachievements that we heard about in the last days

First, after ISO, Spitzer and its mid-infrared spectrometer have considerablywidened the field of PAH studies, especially the extragalactic world up to amaz-ingly large distances close to ten billion light years. This is strikingly demonstratedby the figure displayed by O. Berne showing the rapid increase in papers on PAHsin the last ten years, and furthermore that the number of papers on PAHs inextragalactic systems increased even more rapidly.

The other major trend is the amazing strength of the synergy between astro-physics, laboratory and theory in the works which were reported. Indeed, verylarge efforts have been invested in ambitious, astronomy orientated laboratoryprojects exploiting sophisticated techniques. At the same time, the developmentof a variety of theoretical and computing techniques has spectacularly improvedthe efficiency of quantum chemistry methods, as reviewed by F. Pauzat and others.

Astrophysical PAHs remain highly appealing even outside the strict realm ofastronomy, because of their strong connection with astrochemistry and exotic in-terstellar organic molecules, and also with interstellar dust and thus with the originof planetary systems as traced by the presence of PAHs in meteorites and comets.With SOFIA, JWST and SPICA, exciting new instrumentation will come on lineto probe the many mysteries associated with interstellar PAHs. I will thus list inturn the most salient features of each of the six sessions of the symposium.

2 Rich (IR) spectra of interstellar PAHs

The interpretation of the rich and complex interstellar mid-IR PAH spectra re-mains difficult, even for the main classical features from 6.2 to 12.7 μm, and afortiori the weaker features although many of them are now well established, es-pecially in the 15–20 μm range. However, we have now a very comprehensivegeneral analysis of PAH mid-IR spectra, mainly from the Ames-Dutch group, asreported by E. Peeters. It is based on a rich set of observational data of a broadvariety of astronomical sources, and PAH spectroscopic data that is now availablein the impressive Ames database. The main classes of spectra are well establishedfor various classes of sources including interstellar HII regions, reflection nebulae,circumstellar material of most post-AGB stars, planetary nebulae, Herbig Ae-Bestars, plus peculiar objects. Various characteristics of the spectra reflect first theionization state and the size of the PAHs. In addition, specific interpretations forvarious details have been proposed, involving mainly chemical modifications suchas hetero-atoms - mainly N - metal compounds, carbon isotopes, aliphatic fraction,etc.

PAH band ratios have been further analysed in detail in various observations,and in models e.g. by F. Galliano. An interesting spectral decomposition of PAHemission has been proposed by O. Berne and the Toulouse group. It is based

A. Omont: Summary of the Meeting 465

on the identification of a few main spectral components, attributed to specificcompounds. The addition with various weights of these components may accountfor the spectra observed in different spatial regions.

Although overwhelming, the infrared range is by no way the only spectralrange where PAHs show up in the interstellar medium. I will discuss optical bandsbelow. PAHs have long been proposed to be main contributors to the major bandseen around 2175 A in the UV extinction curve. B. Draine very convincinglyreviewed the arguments in favor of this natural key attribution. As reported by C.Joblin, a major effort is undertaken to detect far-IR modes of PAHs with Herschel.Detection of such modes would represent a major breakthrough because thesemodes should be more specific than in the mid-IR. However, as she discussed, it isnot an easy task. Finally I note the absence in the meeting of a specific discussionof the likely important contribution of PAH rotation to the radio emission of theinterstellar medium around 30 GHz.

3 PAHs, star formation and metallicity

Because of their UV excitation, it is well established that the mid-IR features ofPAHs are major tracers of star formation, as shown by their ubiquitous presencein star formation regions of various types of galaxies. The correlation between8 μm PAH emission and the ionizing photons produced by star formation wascomprehensively and quantitatively reviewed by D. Calzetti, with the conclusionthat it is almost perfect. This emission can thus be used as a star formation rateindicator. However, the 8 μm emission is heavily dependent on metallicity, andthe 8 μm emission may also be triggered by evolved (non star-forming) stellarpopulations.

Detailed evidence for the correlation of PAHs and metallicity was given by sev-eral other communications and posters, including E. Peeters, F. Galliano, L. Huntand K. Sandstrom. This utilizes large samples of galaxies, as well as the SMCwhere the PAH fraction is especially low and spatially variable.

4 The lifecycle of PAHs in space. PAH evolution and processing

PAHs are relatively stable and natural carbonaceous products in hydrogen-richenvironment. However, they are submitted to various destruction processes. Stillthey exist in a diversity of environments including hostile ones . This implies thatthey can form easily by a variety of processes.

PAHs are much faster destroyed in hot plasmas than dust, as seen in the halo ofM 82; this confirms that they are efficiently dissociated in each collision with ener-getic particles, while dust is only slowly sputtered (H.Kaneka). However, PAHs areobserved in a significant fraction of the supernova remnants of the comprehensivestudy reported by J. Rho. Such shocked processed PAHs are strongly correlatedwith carbon dust and shock conditions. PAHs also evolve as a function of the UV


intensity in photo-dissociation regions, possibly in connection with PAH clusters(M. Rapacioli).

As reported by I. Cherchneff, the inner envelope of carbon stars is a natural lo-cation of PAH formation as intermediates to the amorphous carbon dust synthesis.In this context, it is important to note the finding of PAHs and onion-like graphenesheets in meteorite nanograins. Understanding the characteristics of dust forma-tion in carbon star may give interesting clues to the relationship between PAHsand carbonaceous dust products.

However, direct formation and growth of PAHs from large molecules or carbondust in molecular clouds remains very likely a crucial process to explain the ubiq-uity of interstellar PAHs as discussed by P. Woods, W. Geppert and V. Menella.

The frequent detection of PAHs in many circumstellar disks, including mostHerbig Ae stars, also shows their importance in the process of formation of planetsand comet nuclei. In such cases, the well defined geometrical conditions allow aprecise modeling of their interaction with the UV field and their destruction, asdeveloped in sophisticated models by R. Siebenmorgen.

5 Electronic properties

Electronic properties remain a major issue for astronomical PAHs in many re-spects. Research about these properties justify a good part of the major labora-tory projects across the world. A good knowledge of excited state properties isneeded for understanding the complex cascade and relaxation processes resultingin interstellar IR emission induced by UV absorption. Major efforts address thespectroscopy of PAHs in difficult laboratory conditions of cold temperature andgas phase isolated PAHs, relevant for astrophysical conditions. They aim at deter-mining not only spectral characteristics – accurate positions, widths and intensitiesof spectral lines – but also fluorescence and relaxation pathways. As reported byT. Pino, laser spectroscopy is especially useful, including multi-photon processes,induced fluorescence, etc., both in gas phase and matrix isolation spectroscopy.

A major challenge for astrophysical PAHs is that their dominant forms maybe a variety of species that are unstable in usual laboratory conditions – ionized,protonated, deprotonated, dehydrogenated – rather than regular neutral PAHs.P. Sarre especially stressed the importance of protonated and deprotonated PAHs,and their fast reactions, providing calculations of their electronic properties andspectra.

The identification of the carriers of the Diffuse Interstellar Bands (DIBs)is still the main issue in the field of interstellar compounds possibly associated withPAHs. Ted Snow gave a comprehensive summary of the state of the DIB problem.He stressed that while more than 600 DIBs have now been catalogued, with morecoming, not one DIB has been definitively identified, after almost 90 years oftrying! However, almost certainly carbon-based molecules are responsible, andcandidates include linear chains, buckyballs, PAHs, and others. There is some hopethat UV COST/HST spectroscopy gives important clues about this identification.

A. Omont: Summary of the Meeting 467

Nick Cox further reviewed the hypothesis that DIBs are connected with PAHs orother large “related” molecules such as fullerenes, nanotubes or carbon rings.

Some significant progress in this field was reported during the symposium,especially the confirmation of C60 in reflection nebulae through mid-IR detectionof several lines by K. Sellgren.

6 PAHs and carbonaceous grains

The connection of PAHs with carbonaceous grains is obvious from the similitudeof their basic graphene structure, and the possibility to exchange carbonaceousmatter through PAH accretion onto grains and PAH generation in grain shattering.

Furthermore, PAHs could also condensate onto water rich ices in dense clouds.As discussed by Bouwman et al., the UV induced PAH-grain photochemistry maythen play an important role in interstellar chemistry.

However, the relationship of PAHs with carbonaceous grains is much moreintimate because of their similitude of structure and possible family ties. C. Jaegergave a fascinating review of the various solid forms of carbon and carbonaceousparticles. Compared to diamond and even graphite, amorphous hydrogenatedcarbon and soot particles are much more likely akin to PAHs.

Among various topics, she reported Matrix-Assisted-Laser Desorption and Ion-ization combined with mass spectrometry to produce and characterize amazinglyheavy PAHs up to ∼3000 atomic mass units, such as C222H42!

Her summary about soot formation and its multiple processes may give impor-tant clues for interstellar PAHs. She described how different may be soot formationat high temperature, leading to fullerene-like carbon, and at low temperature end-ing with soot with plane layers. It is clear that carbonaceous transition objectsbetween molecules and solids may have an amazing diversity including: PAH clus-ters, PAH oligomers, 3D fragments of fullerenes, oligomers of fullerene fragments,and small nested buckyonions!

7 The role of PAHs in the physics and chemistry of the interstellarmedium

When seeing the important contribution of PAHs in the electromagnetic emissionof the interstellar medium of galaxies, it is not surprizing that they also play acrucial role in its physics and chemistry.

Talks and posters by L. Verstraete, J. Montillaud and others showed thatPAHs intervene in many physical processes. Their determining contribution to theionization of the interstellar gas is the most well known. It is also now agreed thattheir rotation emission may dominate the radio continuum emission of galaxies atwavelengths around 1 cm. However, there are still many pending questions to befurther investigated by observations, modelling, laboratory and theory.

The role of PAHs in the chemistry of the interstellar medium is probably atleast equally important, as discussed in many talks and posters, and especially in


the comprehensive review by V. Bierbaum. It was also stressed by A. Simon thatmetallic compounds, especially Fe-PAH, may be important.

8 Conclusion

To conclude, there is no doubt that the symposium was timely and highly success-ful. The field is blooming and will continue for long. Certainly, the symposium, itsdense scientific program and its wonderful organization by Christine and the localorganizing committee will significantly contribute to a new start of the explorationof galaxies through their PAHs. Exciting new instrumentation is coming for prob-ing their many mysteries, with SOFIA, JWST, SPICA, etc, and one may againanticipate an excellent synergy with the major progress expected in laboratoryand theoretical techniques.

Author Index

Acke B., 259af Ugglas M., 241Allamandola L.J., 109,

251, 305Andersen M., 169

Balog R., 453Baouche S., 453Basire M., 95Bauschlicher C.W., 109Bergin E., 123Bernard J.P., 169Berne O., 49, 223Biennier L., 191Bierbaum V.M., 427Boersma C., 109Bolatto A.D., 215Bot C., 215Boulanger F., 399Bouwman J., 251Brechignac Ph., 95, 355

Calvo F., 95Calzetti D., 133Cami J., 109, 117Candian A., 373Carleton T.M., 209Carpentier Y., 355Cecchi–Pestellini C., 327Chalyavi N., 355Chandrasekaran V., 191Cherchneff I., 177Contreras C.S., 201Cox N.L.J., 349Cuppen H.M., 251

Dartois E., 381Destree J.D., 341

Dopfer O., 103Draine B.T., 29, 215

Evans A., 407

Feraud G., 355Friha H., 355

Galliano F., 43Gehrz R.D., 407Geppert W.D., 241

Hamberg M., 241Hammonds M., 373Helton L.A., 407Henning T., 293Hewitt J., 169Heymann F., 285Hornekær L., 453Hudgins D.M., 109Huisken F., 293Hunt L.K., 143

Im M., 149Ingalls J.G., 209Ishihara D., 157Izotov Y.I., 143

Jager C., 293Joalland B., 223Joblin C., 49, 123, 209,

223, 327, 399, 441, 447Jørgensen B., 453

Kallberg A., 241Kaminska M., 241Kamp I., 271Kaneda H., 55, 157Kashperka I., 241

Kim J.H., 149Klippenstein S.J., 191Kokkin D.L., 355Krugel E., 285

Larsson M., 241Lee H.M., 149Lee M.G., 149Le Page V., 427Linnartz H., 251Luntz A.C., 453

Malloci G., 123, 327Mattioda A.L., 109Maurette M., 319Mennella V., 393Mirtschink A., 223Montillaud J., 223, 447Mouri A., 157Mulas G., 123, 327Mutschke H., 293

Nilsson L., 453

Ohsawa R., 55Okada Y., 55Omont A., 463Onaka T., 55, 157, 399Oomens J., 61

Paal A., 241Parneix P., 95Pathak A., 373Pauzat F., 75Peeters E., 13, 109Pilleri P., 49, 399Pino T., 95, 355Puerta Saborido G., 109

470

Rapacioli M., 223, 441Rauls E., 453Reach W.T., 169Rho J., 169Ricca A., 109Ricketts C.L., 201Rowe B.R., 191

Sabbah H., 191Sakon I., 55, 157Salama F., 201Sanchez de Armas F., 109Sandstrom K.M., 215Sarre P.J., 373Sauvage M., 143Schmidt T.W., 355Sellgren K., 209

Semaniak J., 241Shimonishi T., 55Siebenmorgen R., 285Simon A., 223, 441Simonsson A., 241Sims I.R., 191Smith J.D.T., 209Snow T.P., 341, 427Spiegelman F., 223, 441Stanimirovic S., 215Stensgaard I., 453

Talbi D., 223Tanaka M., 55Tappe A., 169Thomas R.D., 241Thrower J.D., 453

Thuan T.X., 143Tielens A.G.G.M., 3Toublanc D., 447Trippel S., 241Troy T.P., 355

Verstraete L., 415Vigren E., 241

Werner M.W., 209Woods P.M., 235Woodward C.E., 407

Yamagishi M., 157Yasuda A., 157

Zhang M., 241Zhaunerchyk V., 241

Astronomical Object Index

BD+303639 81

3C 396 (G39.2-0.3) 173Ced 201 224Chi Cygni 186Cloverleaf galaxies 7, 8CRL 618 235, 237, 242CRL 2688 382

D7 Cru 409, 410DM Tau 28130 Dor 400, 401, 418

Galactic Center 384,385

HD 10412 273HD 14154 276HD 34282 52HD 95881 273HD 97048 262, 263,

281HD 141569 281HD 163296 281HD 169142 281HD 200775 52

IRAS 11308 173,174

IRAS 08572+3915385

IRAS 13418-6243 120,121

IRAS 23133+6050 120,121

IRC+10216 179, 184,185, 186, 192, 194,195-198, 236, 298,362

IRS 48 281

Jupiter 236

Kes 17 (G304.6+0.1)173, 174

Kes 69 (G21.8-0.6)173, 174

Large MagellanicCloud 17, 20, 48,158, 170, 173,236, 351, 418

Lick Hα 234 173, 174

M17 18, 58, 59, 400,404

M31 58, 351M33 351M81 163M82 17, 35, 52, 53,

161, 162, 1632MASX

J16205879+5425127152

Magellanic clouds 29,30, 31, 40, 137, 162,167, 216ff, 224, 236,351, 403

Milky Way 29, 30

N132D 170, 173

NGC 253 162NGC 628 138NGC 891 161NGC 1068 385NGC 1569 163NGC 2023 20, 210NGC 2974 164NGC 3962 164NGC 4536 15NGC 4589 165, 166NGC 5194 36NGC 5195 35, 36NGC 5529 161NGC 5713 36NGC 5866 36NGC 5907 161NGC 6946 36NGC 7023 50, 51, 85,

210-213, 224, 225,228, 230, 233, 401,403, 450

NGC 7027 4, 6,15, 20,81, 382

NGC 7331 36, 173, 174

Orion Bar 6, 15, 18,20, 58, 127, 128,400, 404

P/Encke 322P/IRAS 322Puppies A supernova

remnant 174, 175

QV Vul 408, 409, 411

RCW 29 160, 161

472

Red Rectangle 120, 121,126, 365

Rho Oph 224RW Leo, see IRC+10216

Saturn 236Serpens molecular cloud

260Small Magellanic Cloud

137, 215ff, 351

SMP LMC 11 236

Titan 236, 242, 247Tycho supernova

remnant 159

V705 Cas 408, 409, 411V842 Cen 408, 409, 411V854 Cen 179, 186, 187

V2361 Cyg 409V2362 Cyg 409, 410V Cygni 186Vela supernova remnant

160VII Zw 353 152

W33A 255Wild-2 322

Chemical Compound Index

Acenaphthene (C12H10)105, 106,205, 206, 430

Acenaphthyl (c-C12H9) 365Acenaphthylene (C12H18) 83, 105,

106, 206Acetylene (C2H2) 184, 185, 192,

193, 195, 196, 201, 205, 206, 237,238, 279, 280

Acridine (C13H9N) 429Ammonia (NH3) 192, 305, 306, 307Anthracene (C14H10) 104-106,

253, 352, 356, 429, 430Azulene (C10H8) 105, 106, 363,

429, 430

Benzene (C6H6) 104-106, 181, 182,184, 186, 194, 227, 235, 236, 237,238, 242ff, 279, 299, 363, 395, 430,432

Benzo(g,h,i)perylene (C22H12) 253Benzene, protonated (C6H+

7 ) 66,103ff, 430

Benzyne (C6H4) 193I,3 butendienyl (l-C4H5) 1821-buten-3-ynyl (l-C3H3) 182

Carbon chain 70Carbondioxide (CO2) 306, 323Carbon monoxide (CO) 185, 186,

279, 305, 306, 307, 320Carbon nanotube 113Carbon sulfide (CS) 185, 192CH+

5 70C3 299, 352, 365C3H2 355C3H4 238C4H 192C4H+

3 237, 238C5 352C6H− 376C78H26 362, 363C96H30 362, 363C222H42 295, 358, 362

Chrysene (C18H12) 180Circumcoronene (C54H18) 195, 443Circumovalene (C66H20) 127, 128CN 192Coranulene (C20H10) 378Coronene (C24H12) 61, 69, 104-106, 113, 114,

194, 195, 227, 233, 253, 357, 364, 376, 377,395, 429, 430, 437, 448, 449, 450, 455, 457,458

coronene, deprotonated (C24H−11) 376

coronene, protonated (C24H+13) 375

Cyclopropenylidene (c-C3H2) 235

Diacetylene (C4H2) 235, 236Dicarbon radical (C2) 187, 299, 344, 365Dihydronaphthalene (C10H10) 436Diamondoids 383

Ethane (C2H6) 203, 204Ethylene (C2H4) 203, 204Ethylene oxide (c-C2H4O) 236Ethylnaphthalene (C22H18) 239Ethynyl radical (C2H) 192, 193, 430

Fluoranthene (C16H10) 83, 84, 429, 430Fluorene (C13H10) 356Formaldehyde (H2CO) 305Formamide (NH2CHO) 305Formyl ion (HCO+) 238Fullerene (C60 and others) 209ff, 297,

301, 302, 351, 353, 383, 429, 431, 432Fulvene (C6H6) 181

Graphene 113Graphite 30

H2 244, 247, 276, 321, 322, 394, 419HC4H+ 352HC11N 305, 361Hexa-peri-hexabenzocoronene (C42H18)

362, 374Hydrogen cyanide (HCN) 185, 192, 320Hydronaphthalene (C10H9) 365Hydronaphthyl (C10H9) 431

474

Hydronium (H3O+) 70Hydroxyl (OH) 186, 430

Indanyl (C9H9) 365Indynyl (C9H7) 365

Methane (CH4) 203, 204, 279, 280,306, 320

Methanol (CH3OH) 306, 307Methyl (CH3) 430Methylene (CH2) 182Methylidyne (CH) 349, 353, 365, 430Methylidyne cation (CH+) 349Methylnaphthalene (C11H10) 205

Naphthalene (C10H8) 64, 65, 66, 69,81, 82, 86, 88, 104, 106, 126, 127,203, 205, 231, 239, 352, 359, 363,364, 428, 429, 430, 432, 449

Naphthalene, protonated (C10H+9 )

104-106Naphthylmethyl (C11H9) 365

Ovalene (C32H14) 362

Pentacene (C22H14) 105, 106, 429Perhydrocoronene (C24H36) 457Perylene (C20H12) 105, 106, 363, 429

Phenanthrene (C14H10) 63, 180, 429,430

Phenyl (C6H5) 181, 186, 239Phylpropargyl (C9H7) 365Polyynes 182, 186, 187, 294, 301Propargyl (C3H3) 181, 182, 186, 193, 238Propynylidene (C3H) 192Pyrene (C16H10) 68, 69, 97, 98, 105, 106,

113, 127, 182, 193, 194, 195, 198, 252,253, 312, 363, 428, 430, 432, 443

Pyrimidine (C4H4N2) 309

Quaterrylene (C40H20) 310

Silicon monoxide (SiO) 186Sulfur dioxide (SO2) 323

Terrylene (C30H16) 362Tetracene (C18H12) 105, 106, 429Triacetylene (C6H2) 235, 236Trihydronaphthalene (C10H11) 365Triphenylene (C18H12) 362, 429

Uracil (C4H2N2O2) 309

Vinylacetylene (C4H4) 182

Water (H2O) 186, 306, 307, 320, 323

Subject Index

Absorption feature, aliphatics 384,394

Active Galactic Nuclei, see Galaxies,AGNAKARI 49ff, 143ff, 401Aliphatic hydrocarbons 23, 266, 384,

387, 395, 396, 402, 403, 408Amino acids 309Amorphous carbon 178, 184, 296,

382, 409, 411Amphilic compound 308Anharmonicity 68, 80, 86, 87, 95ff,

107, 119, 125, 444Anomalous radio emission 30, 33,

124, 274, 419Aromatic carbon 30, 329, 394, 408Astrobiology 305Asymptotic Giant Branch stars 46,

137, 178, 179, 180, 192, 194, 210,211, 216, 218, 219, 236, 279, 295,298, 382

Astration 57

Band gap 294, 295Bay region 24Blue Compact Dwarfs, see Galaxies,Blue Compact DwarfBig Grains (BG) 173

Canonical ensemble 88, 125, 127Carbide 192Carbonaceous material 31, 46, 47,

294ffCarbonate 323Carbon chain 70, 201, 352Carbon dust 294ffCarbon onion 297, 301, 302Carbon star, see Asymptotic GiantBranch starsCavity ring-down spectroscopy 361C-C modes 18, 80, 82, 99, 107C-H modes 17, 18, 23, 66, 80, 82, 99,

105, 107

Chemistry models 183-186, 237-239,279, 433-435, 448-451

Chemistry, non-equilibrium 185, 186Cloverleaf galaxies 7, 8Cluster, see also PAH, clustersCluster, dissociation energy 227, 229Cluster, ionization energy 229Cluster, structure 227Coal 5, 297Combustion 178, 236, 298Comets 307, 312, 314, 315, 322Complexes, see PAH, complexesCosmic infrared background 8Cosmic Rays 324, 325

Database 84, 109ff, 122ffDehydrogenation, see PAH, dehydro-genationDensity Functional Theory (DFT)

79, 110, 118, 129, 225, 333, 442-444Density-of-states 96Depletion 30, 404Destruction, see PAH, destruction &dust, destructionDeuterium 57, 403Diamond 294, 301, 336Diamondoids 383Diffuse Interstellar Bands (DIBs) 5,

123, 187, 209, 328, 331, 341ff, 356,373, 374

Diffuse Interstellar Bands, families343, 344

Diffuse Interstellar Bands,polarization 351

Diffuse Interstellar Bands, properties343, 344

Diffuse Interstellar Medium 19,416-418

Dimer, see PAH, dimerDipole moment 378Dissociative recombination 242ffDrumhead modes 62, 111-113Dust 29ff, 160

476

Dust, annealing 320Dust, carbon 294ff, 329, 381ff, 394ffDust, destruction 170, 173Dust, evolution 46, 47, 170, 173, 415Dust, formation 177ff, 192, 197, 198,

201, 202, 212, 236, 365, 382, 408,410

Dust, H2 formation 321, 322, 394,395

Dust, IR emission 30, 32, 34-36, 134Dust, lifecycle 381Dust, Lunar 319, 320Dust, models 29ff, 43ff, 285, 286, 331,

332Dust, nova 407ffDust, shattering 158, 164, 216, 236Dust, size distribution 31, 32, 158,

164, 172, 173, 286, 415Dust, structure 294, 2950297Dust, surface reactions 238, 393ffDust-to-gas ratio 45Dwarf galaxy, see galaxies, dwarf

Electron affinity, see PAH, electronaffinity

Electron attachment 429Electron correlation 80Electronic spectroscopy 327ff, 355ff,

374ffElectron recombination 429, 430,

434, 437Elliptical galaxy, see Galaxies,

ellipticalEmission model, see PAH, emission

modelExtinction 29Extinction curve 328ffExtinction bump (2175A) 30, 31, 39,

40, 328-330, 331, 332, 351, 362, 418Extinction, far-UV 331, 332, 334, 418Evolved star, see Asymptotic GiantBranch star

Far-infrared absorption 111-113Fluorescence 359, 360, 365Free electron laser 67, 69Fullerene 209ff, 297, 301, 302, 351,

353, 383, 429, 431, 432

FU Orionis star 261Fused silica 322

Galactic halo 161, 161Galactic plane 57Galactic superwind 158, 161, 162,

163Galaxies 7, 19Galaxies, AGN 152, 164, 165, 384Galaxies, Blue Compact Dwarf 144Galaxies, dwarf 46, 137, 144, 163Galaxies, elliptical 164Galaxies, evolution 46, 152Galaxies, high redshift 133, 134Galaxies, LINERS 164, 165Galaxies, low metallicity 216Galaxies, star formation 134Galaxies, starburst 7, 30, 39, 143,150, 152, 153, 387Galaxies, submillimeter 7Gas, cooling 417, 423Gas, ionization 418Gas phase spectroscopy 63Giant planets 236, 242Graphane 295Graphene 113, 180, 294, 295, 296,

299Graphite 30, 179, 294, 322, 330

H2 formation 244, 247, 276, 321, 322,394, 395, 419, 435, 436

HAC, see Hydrogenated amorphousCarbonHACA mechanism 182, 186, 187, 193Hartree-Fock method 79Herbig AeBe star 17, 19, 20, 260, 261,262, 263, 264, 272, 273, 276, 278, 383Herschel Space Observatory 127Heteroatom substitution 22HII region 17, 19, 147Hot plasma 158, 163Hydrogenated Amorphous Carbon 4,

5, 193, 212, 296, 297, 384-387, 394,403, 408

Ice, interstellar 251, 280, 306Ice, photochemistry 307-312Ilmenite 321

477

Infrared cirrus 5, 7Infrared emission spectroscopy 63, 68Infrared features, emission plateau

16Infrared features, peak position 18,

272Infrared features, profile 18, 85, 98,99, 114, 119, 125, 128, 145, 265, 272,

387Infrared features, relative intensities

16-18, 144Infrared features, spatial variations

16-18Infrared features, substructure 15Infrared features, variations 16-18,

35, 37, 56, 57, 121, 442, 443Infrared Multiphoton DissociationSpectroscopy 66, 67, 69, 105ff, 360,

361Infrared spectroscopy 61ff, 251ffInternal Conversion (IC) 366, 367Interplanetary dust particle 307, 314,

315InterSystem Crossing (ISC) 366, 367Ion irradiation 320, 321Ionization parameter 50Ion-neutral reactions 431, 437IRAS LRS 6ISOCAM 7ISO SWS 7Intramolecular Vibrational Redistri-bution 67, 96, 125, 356, 357, 366, 367

Kerogen 322-325

Laboratory method 61ff, 103ff,203-205, 241ff, 251ff, 427ff,357-361

Magnesium Sulfide 184Matrix isolation spectroscopy 62, 69,

106, 107, 252-255Membrane 308Meteorites 179, 307, 309, 314, 315,

322, 323, 383, 384Messenger-atom spectroscopy 65, 66,

69, 361Metallicity 17, 18, 43, 135-137, 143,

163, 216Methylpolyynes 235

Microcanonical ensemble 88, 97, 125,127

Micrometeorite 322, 323Molecular beam 68Molecular cloud 218Molecular ions 64-68Molecular structure, see PAH, molec-ular structure

Nanodiamonds 336, 383Nanotubes 113, 294, 295Nova 407ff

Olivine 321oop modes 17, 18, 23Origin of life 314, 315

PAH, abundance 13, 43ff, 135, 217,271, 273, 330, 416

PAH, anion 429, 43§1, 437PAH, cation 122, 267PAH, charge balance 23, 24, 31, 34,

37PAH, chemical model 183-186, 237-

239, 279, 433-435, 448-451PAH, chemistry 177ff, 191ff, 201ff,

235ff, 251ff, 278, 427ffPAH, chemical evolution 50PAH, chemical growth 181PAH, clusters 21, 22, 104, 182, 186,

194, 198, 266, 301, 388, 441ffPAH, complexes 22, 62, 67, 70, 82,

388, 441ffPAH, database 84, 109ff, 117ffPAH, dehydrogenation 23, 76, 80, 82,

97, 98, 99, 242, 275, 276, 279, 332,333, 334, 351, 434, 436, 442, 448,449, 450

PAH, deprotonated 373PAH, destruction 43, 48, 137, 157ff,

170, 173, 216, 289PAH, deuteration 57-59, 82, 403PAH, dimer 183, 194, 195, 298, 301PAH, dissociation 67, 428, 434, 447,

449PAH, dissociation threshold 67PAH, dissociative recombination

242ffPAH, electron affinity 376, 421, 429PAH, electronic transition 124

478

PAH, emission model 29ff, 88, 98, 99,118, 119, 124, 125

PAH, evolution 46, 157ffPAH, family 22PAH, far-infrared 109ff, 123ffPAH, formation 45, 137, 177ff, 191-

194, 201, 216, 218, 236, 279, 295,298, 299

PAH, fragmentation 428PAH, H2 formation 244, 247, 276,

419, 435, 436, 453ffPAH, hydrogenated 82PAH, hypothesis 3ffPAH, ice 251ff, 261, 309-311PAH, identification 111ff, 123ff,

341ff, 349ffPAH, infrared cascade 97PAH, infrared spectroscopy 61ff, 75ff,

109ff, 441ffPAH, ionization 23, 31, 34, 37, 50,

51, 164, 219, 275, 276, 344, 351,353, 420, 428, 434, 448

PAH, metal complex 22, 62, 67, 70,82, 388, 441ff

PAH, molecular structure 23, 24PAH, negative ion 82, 122,PAH, neutral 122PAH, nitrogen substitution 22, 62,

67, 110, 118, 122, 388PAH, photochemistry 253, 254, 289,

428, 447, 449PAH, photo-detachment 360PAH, photo-ionization 360PAH, photolysis 251PAH, protonated 67, 70, 103ff, 242,

245-247, 373, 432PAH, reaction 427ff, 449PAH, recombination 448PAH, rotational emission 33, 420PAH, rotational excitation 61, 127,

377PAH, sidegroup 309PAH, size 22, 23, 32, 34, 39, 146, 260,

267, 275PAH, spectral classes 18-20, 111, 121PAH, spectral variations 14, 16-20,

144, 442, 443PAH, sputtering 148, 158

PAH, superhydrogenated 431, 435,436, 448, 453ff

PAH, temperature 32, 33PAH, UV feature (2175A) 39, 327ffPAH, UV processing 20, 51PAH, UV spectrum 330ffPAH, vibrational modes 14, 15PAHfit 144PANHs 22, 62, 67

Photochemistry, see PAH, photo-chemistryPhotoDissociation Region (PDR)

224, 228, 229, 266, 400, 422, 423,448, 450, 457

Photo-detachment 360, 429Photo-electric heating 280, 416Photo-ionization 360, 416, 421Planetary nebula 17, 19, 24, 193, 224,

266Plasma discharge 203-205Platt particles 329, 330Polarization 30, 357Post-AGB object 19, 20, 46, 193, 210,

236Prebiotic molecules 307-309, 314, 315Presolar grains 30, 179, 194Protoplanetary disk 238, 260, 261,

271Protoplanetary disk, models 278, 280Protoplanetary disk, structure 263-

365, 273, 281, 287, 288Protoplanetary disk, sedimentation

263, 264, 278, 279, 281Pyrolysis 180

Quantum chemistry 75ff, 333, 454Quenched Carbonaceous Compound

(QCC) 5, 297

Radiation damage 320Radiation field, hardness 17, 18, 36,

37, 39, 137, 138, 148, 164, 165, 262,266, 289

Radiation field, intensity 148, 172,180, 281, 352

Radiation field, interstellar 35, 39Radiative transfer 286Radical reactions 430

479

RCrB stars 178, 179, 186, 210, 211,213

Reflection nebula 17, 19Rotational spectroscopy 124, 377,

378Ro-vibrational profile 127

Sapphire 321Scattering 30, 170, 172SED, see Spectral Energy

DistributionShock 158, 170, 171, 183-186, 192,

194, 216, 298Silica 321Silicates 30, 31, 411Silicon carbide 184, 295Sodalime glass 322Spectral decomposition 21Spectral Energy Distribution 217,

273, 274, 287Spinning dust 419Spitzer IRS 7Sputtering 170, 172S Star 180, 181Starburst galaxy, see Galaxies,

starburstStar formation history 45Star formation rate 14, 24, 133ff, 150Stellar ejecta 178, 192Stellar population model 45Stellar pulsation 185Stellar wind 178, 183-186, 195-197,

237

Submillimeter galaxy, see Galaxies,submillimeterSulfide 323Supernova 45, 137, 178, 179Supernova remnant 158, 169ff

Tauc relation 297Temperature, radiative equilibrium 4Thermal approximation, see canoni-cal ensembleThermochemistry 433T-Tauri star 19, 20, 260, 261, 264,

272, 276, 289

UIR bands 4, 6,7, 75UV absorption 350UV bump (2175A) 30, 31, 39, 40,

328-330, 331, 332, 351, 356, 362,418

UV extinction 39, 345, 346, 418

Very Small Grains (VSG) 18, 21, 50,172, 173, 223ff, 266, 274, 275, 280,403, 442

Very Small Grains, evaporating 50,228

Vesicles 308

Wolf Rayet stars 178, 179, 211

X-ray 261, 410

Zodiacal cloud 322

EAS Publications Series

Volume 1 AGN in their Cosmic Environment (JENAM’99)

September 7-9, 1999 • 30 e

Volume 2 GAIA: A European Space Project

May 14-18, 2001 • 60 e

Volume 3 Star Formation and the Physics of Young Stars

(Summer School on Stellar Physics X)

September 18-22, 2000 • 40 e

Volume 4 Infrared and Submillimeter Space Astronomy

June 11-13, 2001 • 65 e

Volume 5 Radiative Transfer and Hydrodynamics in Astrophysics (GRETA)

November 28, 2001 • 25 e

Volume 6 Observing with the VLT Interferometer (EuroWinter School)

February 3-8, 2002 • 47 e

Volume 7 Final Stages of Stellar Evolution (Summer School on Stellar Physics XI)

September 16-21, 2001 • 52 e

Volume 8 Astronomy with High Contrast Imaging

May 13-16, 2002 • 56 e

Volume 9 Magnetism and Activity of the Sun and Stars

September 17-21, 2002 • 56 e

Volume 10 Galactic & Stellar Dynamics (JENAM 2002)

September 3-6, 2002 • 36 e

Volume 11 The Future Astronuclear Physics

August 20-22, 2003 • 47.50 e

Volume 12 Astronomy with High Contrast Imaging II

October 6-10, 2003 • 56 e

Volume 13 Evolution of Massive Stars, Mass Loss and Winds

(Summer School on Stellar Physics XII-XIII)

October 13-18, 2002 and October 6-10, 2003 • 50 e

Volume 14 Dome C: Astronomy and Astrophysics Meeting

June 28-30, 2004 • 63 e

Volume 15 Radio Astronomy from Karl Jansky to Microjansky (JENAM 2003)

August 27-30, 2003 • 72 e

Volume 16 Teaching and Communicating Astronomy (JENAM 2004)

September 13-17, 2004 • 44 e

Volume 17 Element Stratification in Stars: 40 Years of Atomic Diffusion

Meeting in Honour of Georges Michaud

June 6-10, 2005 • 60 e

Volume 18 Radiative Transfer and Applications to Very Large Telescopes

(GRETA)

May 11-13, 2005 • 47 e

Volume 19 Stars and Nuclei: A Tribute to Manuel Forestini

March 4-5, 2004 • 36 e

Volume 20 Mass Profiles and Shapes of Cosmological Structures (IAP 2005)

July 4-9, 2005 • 47 e

Volume 21 Stellar Fluid Dynamics and Numerical Simulations:

From the Sun to Neutron Stars

September, 2004 and May, 2005 • 55 eVolume 22 Astronomy with High Contrast Imaging III: Instrumental Techniques,

Modeling and data Processing

May 16-19, 2005 • 55 e

Volume 23 Sky Polarisation at Far-Infrared to Radio Wavelengths: The Galactic

Screen before the Cosmic Microwave Background

September 12-15, 2005 • 39 e

Volume 24 CRAL-2006. Chemodynamics: From First Starts to Local Galaxies

July 10-14, 2006 • 47 e

Volume 25 1st ARENA Conference on “Large Astronomical Infrastructures at

CONCORDIA, prospects and constraints for Antarctic Optical/IR

Astronomy”

October 16-19, 2006 • 47 e

Volume 26 Stellar Evolution and Seismic Tools for Asteroseismology

November 20-23, 2006 • 32 e

Volume 27 The Third European Summer School on Experimental Nuclear

Astrophysics

October 2-9, 2005 • 39 e

Volume 28 Perspectives in Radiative Transfer and Interferometry

May 14-16, 2007 • 25 e

Volume 29 Tidal Effects in stars, Planets and Disks

September 5-9, 2005 • 47 e

Volume 30 Spanish Relativity Meeting – Encuentros Relativistas Espanoles

ERE2007

September 10-14, 2007 • 63 eVolume 31 Far-Infrared Workshop 2007

November 5-7, 2007 • 39 e

Volume 32 Stellar Nucleosynthesis 50 Years after B2FH

December 4-8, 2006 • 63 e

Volume 33 2nd ARENA Conference on “The Astrophysical Science Cases at Dome C”

September 17-21, 2007 • 55 e

Volume 34 Astronomy in the Submillimeter and Far Infrared Domains

with the Herschel Space Observatory

April 23-May 4, 2007 • 47 e

Volume 35 Interstellar Dust from Astronomical Observations to Fundamental Studies

May 1st-5, 2006 • 47 e

Volume 36 Dark Energy and Dark Matter

and Theories

July 7-11, 2008 • 55 e

Volume 37 Astrophysics Detector Workshop 2008

November 17-20, 2008 • 63 e

Volume 38 Nonlinear Pulsations and Hydrodynamics of Cepheids

July 9-13, 2007 • 32 e

Volume 39 Stellar Magnetism

September 24-28, 2007 • 39 e

Volume 40 3rd ARENA Conference on “An Astronomical Observatory at CONCORDIA

(Dome C, Antarctica)”

May 11-15, 2009 • 87 e

Volume 41 Physics and Astrophysics of Planetary Systems

February 18-29, 2008 • 95 e

Volume 42 Extrasolar Planets in Multi-Body Systems: Theory and Observations

August 25-29, 2008 • 79 e

Volume 43 Non-LTE Line Formation for Trace Elements in Stellar Atmospheres

July 30 – August 03, 2007 • 43 e

Volume 44 JENAM 2008: Grand Challenges in Computational Astrophysics

September 8-11, 2008 • 26e

Volume 45 GAIA: At the Frontiers of Astrometry

June 7-11, 2010 • 78e

Order directly to EDP Sciences

17 av. du Hoggar • BP. 112 • P.A. de Courtabœuf • 91944 Les Ulis Cedex A • France

Tel.: 33 (0)1 69 18 75 75 • Fax: 33 (0)1 69 86 06 78 • [email protected]

A Symposium to Celebrate the 25th Anniversary of the PAH ...

Documents

Transcript of A Symposium to Celebrate the 25th Anniversary of the PAH ...